0.2.58

Merge pull request #2661 from langchain-ai/nc/5dec/perf
lib: Performance improvements
2026-08-18 05:35:43 +02:00 · 2024-12-10 14:09:11 -08:00 · 2024-12-10 14:04:02 -08:00 · 2024-12-11 02:00:21 +04:00 · 2024-12-10 13:56:26 -08:00 · 2024-12-10 22:52:51 +01:00
170 changed files with 16639 additions and 6360 deletions
@@ -7,35 +7,29 @@ body:
      value: >
        Thank you for taking the time to file a bug report. 
        
-        Use this to report bugs in LangChain. 
-        
-        If you're not certain that your issue is due to a bug in LangChain, please use [GitHub Discussions](https://github.com/langchain-ai/langchain/discussions)
-        to ask for help with your issue.
+        Use this to report BUGS in LangGraph. For usage questions, feature requests and general design questions, please use [GitHub Discussions](https://github.com/langchain-ai/langgraph/discussions).
        
        Relevant links to check before filing a bug report to see if your issue has already been reported, fixed or
        if there's another way to solve your problem:
        
-        [LangGraph documentation](https://langchain-ai.github.io/langgraph/).
+        [LangGraph Github Discussions](https://github.com/langchain-ai/langgraph/discussions),
+        [LangGraph Github Issues](https://github.com/langchain-ai/langgraph/issues),
+        [LangGraph how-to guides](https://langchain-ai.github.io/langgraph/how-tos/).
        [LangChain documentation with the integrated search](https://python.langchain.com/docs/get_started/introduction),
        [GitHub search](https://github.com/langchain-ai/langgraph),
-        [LangChain Github Discussions](https://github.com/langchain-ai/langgraph/discussions),
-        [LangChain Github Issues](https://github.com/langchain-ai/langgraph/issues),
-        [LangChain ChatBot](https://chat.langchain.com/)
  - type: checkboxes
    id: checks
    attributes:
      label: Checked other resources
-      description: Please confirm and check all the following options.
+      description: Before submitting this issue, please confirm that you have completed all the steps below by checking each option. These steps help ensure your issue is well-defined, relevant, and actionable.
      options:
-        - label: I added a very descriptive title to this issue.
+        - label: This is a bug, not a usage question. For questions, please use GitHub Discussions.
          required: true
-        - label: I searched the [LangGraph](https://langchain-ai.github.io/langgraph/)/LangChain documentation with the integrated search.
+        - label: I added a clear and detailed title that summarizes the issue.
          required: true
-        - label: I used the GitHub search to find a similar question and didn't find it.
+        - label: I read what a minimal reproducible example is (https://stackoverflow.com/help/minimal-reproducible-example).
          required: true
-        - label: I am sure that this is a bug in LangGraph/LangChain rather than my code.
-          required: true
-        - label: I am sure this is better as an issue [rather than a GitHub discussion](https://github.com/langchain-ai/langgraph/discussions/new/choose), since this is a LangGraph bug and not a design question.
+        - label: I included a self-contained, minimal example that demonstrates the issue INCLUDING all the relevant imports. The code run AS IS to reproduce the issue.
          required: true
  - type: textarea
    id: reproduction
@@ -45,14 +39,6 @@ body:
      label: Example Code
      description: |
        Please add a self-contained, [minimal, reproducible, example](https://stackoverflow.com/help/minimal-reproducible-example) with your use case.
-        
-        If a maintainer can copy it, run it, and see it right away, there's a much higher chance that you'll be able to get help.
-        
-        **Important!** 
-        
-        * Reduce your code to the minimum required to reproduce the issue if possible. This makes it much easier for others to help you.
-        * Avoid screenshots when possible, as they are hard to read and (more importantly) don't allow others to copy-and-paste your code.
-
      placeholder: |
        from langgraph.graph import StateGraph

@@ -92,25 +78,8 @@ body:
    attributes:
      label: System Info
      description: |
-        Please share your system info with us. 
-        
-        "pip freeze | grep langchain" 
-        platform (windows / linux / mac)
-        python version
-        
-        OR if you're on a recent version of langchain-core you can paste the output of:
-        
        python -m langchain_core.sys_info
      placeholder: |
-        "pip freeze | grep langgraph"
-        platform
-        python version
-        
-        Alternatively, if you're on a recent version of langchain-core you can paste the output of:
-        
        python -m langchain_core.sys_info
-        
-        These will only surface LangChain packages, don't forget to include any other relevant
-        packages you're using (if you're not sure what's relevant, you can paste the entire output of `pip freeze`).
    validations:
      required: true
@@ -22,8 +22,7 @@ def test(
        # check docker available
        capabilities = langgraph_cli.docker.check_capabilities(runner)
        # open config
-        with open(config) as f:
-            config_json = langgraph_cli.config.validate_config(json.load(f))
+        config_json = langgraph_cli.config.validate_config_file(config)

        set("Running...")
        args = [
@@ -42,7 +42,6 @@ jobs:
        with:
          python-version: ${{ matrix.python-version }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: lint-${{ inputs.working-directory }}

      - name: Check Poetry File
@@ -31,7 +31,6 @@ jobs:
        with:
          python-version: ${{ matrix.python-version }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: test-${{ inputs.working-directory }}
      - name: Login to Docker Hub
        uses: docker/login-action@v3
@@ -60,7 +60,7 @@ jobs:
        env:
          LANGGRAPH_FF_SEND_V2: ${{ matrix.ff-send-v2 }}
        run: |
-          make test
+          make test_parallel

      - name: Ensure the tests did not create any additional files
        shell: bash
@@ -29,7 +29,6 @@ jobs:
        with:
          python-version: ${{ env.PYTHON_VERSION }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: release

      # We want to keep this build stage *separate* from the release stage,
@@ -88,6 +88,7 @@ jobs:
              --check-links-ignore "https://(api|web|docs)\.smith\.langchain\.com/.*" \
              --check-links-ignore "https://x.com/.*" \
              --check-links-ignore "https://github\.com/.*" \
+              --check-links-ignore "http://localhost:8123/.*" \
              --check-links-ignore "/.*\.(ipynb|html)$" \
              --check-links-ignore "https://python\.langchain\.com/.*" \
              --check-links-ignore "https://openai\.com/.*" \
@@ -104,6 +105,7 @@ jobs:
              echo "Running link check on HTML files matching changed notebook files..."
              poetry run pytest -v \
                --check-links-ignore "https://(api|web|docs)\.smith\.langchain\.com/.*" \
+                --check-links-ignore "http://localhost:8123/.*" \
                --check-links-ignore "https://x.com/.*" \
                --check-links-ignore "https://github\.com/.*" \
                --check-links-ignore "/.*\.(ipynb|html)$" \
@@ -31,7 +31,6 @@ jobs:
        with:
          python-version: ${{ env.PYTHON_VERSION }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: release

      # We want to keep this build stage *separate* from the release stage,
@@ -169,7 +168,6 @@ jobs:
        with:
          python-version: ${{ env.PYTHON_VERSION }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}

      - name: Import published package
        shell: bash
@@ -256,7 +254,6 @@ jobs:
        with:
          python-version: ${{ env.PYTHON_VERSION }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: release

      - uses: actions/download-artifact@v4
@@ -298,7 +295,6 @@ jobs:
        with:
          python-version: ${{ env.PYTHON_VERSION }}
          poetry-version: ${{ env.POETRY_VERSION }}
-          working-directory: ${{ inputs.working-directory }}
          cache-key: release

      - uses: actions/download-artifact@v4
@@ -49,7 +49,7 @@ gain understanding of concepts and how they interact by showing one way to achie

 They should **avoid** giving
 multiple permutations of ways to achieve that goal in-depth. Choice is burdensome. Instead, they should guide a new user through a recommended path to accomplishing a concrete goal. While the end result of a tutorial does not necessarily need to
-be completely production-ready, it should be useful and practically satisfy the the goal that you clearly stated in the tutorial's introduction.
+be completely production-ready, it should be useful and practically satisfy the goal that you clearly stated in the tutorial's introduction.

 To quote the Diataxis website:

@@ -13,7 +13,7 @@ serve-clean-docs: clean-docs
 	poetry run python -m mkdocs serve -c -f docs/mkdocs.yml --strict -w ./libs/langgraph

 serve-docs: build-typedoc
-	poetry run python -m mkdocs serve -f docs/mkdocs.yml -w ./libs/langgraph --dirty
+	poetry run python -m mkdocs serve -f docs/mkdocs.yml -w ./libs/langgraph  -w ./libs/checkpoint --dirty

 clean-docs:
 	find ./docs/docs -name "*.ipynb" -type f -delete
@@ -238,7 +238,7 @@ final_state["messages"][-1].content
 * [How-to Guides](https://langchain-ai.github.io/langgraph/how-tos/): Accomplish specific things within LangGraph, from streaming, to adding memory & persistence, to common design patterns (branching, subgraphs, etc.), these are the place to go if you want to copy and run a specific code snippet.
 * [Conceptual Guides](https://langchain-ai.github.io/langgraph/concepts/high_level/): In-depth explanations of the key concepts and principles behind LangGraph, such as nodes, edges, state and more.
 * [API Reference](https://langchain-ai.github.io/langgraph/reference/graphs/): Review important classes and methods, simple examples of how to use the graph and checkpointing APIs, higher-level prebuilt components and more.
-* [Cloud (beta)](https://langchain-ai.github.io/langgraph/cloud/): With one click, deploy LangGraph applications to LangGraph Cloud.
+* [LangGraph Platform](https://langchain-ai.github.io/langgraph/concepts/#langgraph-platform): LangGraph Platform is a commercial solution for deploying agentic applications in production, built on the open-source LangGraph framework.

 ## Contributing

@@ -36,10 +36,11 @@ NOTEBOOKS_NO_EXECUTION = [
    "docs/docs/tutorials/rag/langgraph_self_rag_local.ipynb",
    # this loads a massive dataset from gcp
    "docs/docs/tutorials/usaco/usaco.ipynb",
+    # TODO: figure out why autogen notebook is not runnable (they are just hanging. possible due to code execution?)
+    "docs/docs/how-tos/autogen-integration.ipynb",
    # TODO: need to update these notebooks to make sure they are runnable in CI
    "docs/docs/tutorials/storm/storm.ipynb",  # issues only when running with VCR
    "docs/docs/tutorials/lats/lats.ipynb",  # issues only when running with VCR
-    "docs/docs/tutorials/multi_agent/hierarchical_agent_teams.ipynb",  # taking a very long time to run
    "docs/docs/tutorials/rag/langgraph_crag.ipynb",  # flakiness from tavily
    "docs/docs/tutorials/rag/langgraph_adaptive_rag.ipynb",  # Cannot create a consistent method resolution error from VCR
    "docs/docs/how-tos/map-reduce.ipynb"  # flakiness from structured output, only when running with VCR
@@ -0,0 +1,123 @@
+# How to add semantic search to your LangGraph deployment
+
+This guide explains how to add semantic search to your LangGraph deployment's cross-thread [store](../../concepts/persistence.md#memory-store), so that your agent can search for memories and other documents by semantic similarity.
+
+## Prerequisites
+
+- A LangGraph deployment (see [how to deploy](setup_pyproject.md))
+- API keys for your embedding provider (in this case, OpenAI)
+- `langchain >= 0.3.8` (if you specify using the string format below)
+
+## Steps
+
+1. Update your `langgraph.json` configuration file to include the store configuration:
+
+```json
+{
+    ...
+    "store": {
+        "index": {
+            "embed": "openai:text-embeddings-3-small",
+            "dims": 1536,
+            "fields": ["$"]
+        }
+    }
+}
+```
+
+This configuration:
+
+- Uses OpenAI's text-embeddings-3-small model for generating embeddings
+- Sets the embedding dimension to 1536 (matching the model's output)
+- Indexes all fields in your stored data (`["$"]` means index everything, or specify specific fields like `["text", "metadata.title"]`)
+
+2. To use the string embedding format above, make sure your dependencies include `langchain >= 0.3.8`:
+
+```toml
+# In pyproject.toml
+[project]
+dependencies = [
+    "langchain>=0.3.8"
+]
+```
+
+Or if using requirements.txt:
+
+```
+langchain>=0.3.8
+```
+
+## Usage
+
+Once configured, you can use semantic search in your LangGraph nodes. The store requires a namespace tuple to organize memories:
+
+```python
+def search_memory(state: State, *, store: BaseStore):
+    # Search the store using semantic similarity
+    # The namespace tuple helps organize different types of memories
+    # e.g., ("user_facts", "preferences") or ("conversation", "summaries")
+    results = store.search(
+        namespace=("memory", "facts"),  # Organize memories by type
+        query="your search query",
+        limit=3  # number of results to return
+    )
+    return results
+```
+
+## Custom Embeddings
+
+If you want to use custom embeddings, you can pass a path to a custom embedding function:
+
+```json
+{
+    ...
+    "store": {
+        "index": {
+            "embed": "path/to/embedding_function.py:embed",
+            "dims": 1536,
+            "fields": ["$"]
+        }
+    }
+}
+```
+
+The deployment will look for the function in the specified path. The function must be async and accept a list of strings:
+
+```python
+# path/to/embedding_function.py
+from openai import AsyncOpenAI
+
+client = AsyncOpenAI()
+
+async def aembed_texts(texts: list[str]) -> list[list[float]]:
+    """Custom embedding function that must:
+    1. Be async
+    2. Accept a list of strings
+    3. Return a list of float arrays (embeddings)
+    """
+    response = await client.embeddings.create(
+        model="text-embedding-3-small",
+        input=texts
+    )
+    return [e.embedding for e in response.data]
+```
+
+## Querying via the API
+
+You can also query the store using the LangGraph SDK. Since the SDK uses async operations:
+
+```python
+from langgraph_sdk import get_client
+
+async def search_store():
+    client = get_client()
+    results = await client.store.search_items(
+        ("memory", "facts"),
+        query="your search query",
+        limit=3  # number of results to return
+    )
+    return results
+
+# Use in an async context
+results = await search_store()
+```
@@ -36,8 +36,8 @@ Dependencies can optionally be specified in one of the following files: `pyproje
 The dependencies below will be included in the image, you can also use them in your code, as long as with a compatible version range:

 ```
-langgraph>=0.2.30,<0.3.0
-langgraph-checkpoint>=1.0.14
+langgraph>=0.2.56,<0.3.0
+langgraph-checkpoint>=2.0.5,<3.0
 langchain-core>=0.2.38,<0.4.0
 langsmith>=0.1.63
 orjson>=3.9.7
@@ -36,8 +36,8 @@ Dependencies can optionally be specified in one of the following files: `pyproje
 The dependencies below will be included in the image, you can also use them in your code, as long as with a compatible version range:

 ```
-langgraph>=0.2.30,<0.3.0
-langgraph-checkpoint>=1.0.14
+langgraph>=0.2.56,<0.3.0
+langgraph-checkpoint>=2.0.5,<3.0
 langchain-core>=0.2.38,<0.4.0
 langsmith>=0.1.63
 orjson>=3.9.7
@@ -21,7 +21,7 @@ Install the proper packages:
 Ensure you have an API key, which you can create from the [LangSmith UI](https://smith.langchain.com) (Settings > API Keys). This is required to authenticate that you have LangGraph Cloud access. After you have saved the key to a safe place, place the following line in your `.env` file:

 ```python
-LANGCHAIN_API_KEY = *********
+LANGSMITH_API_KEY = *********
 ```

 ## Start the API server
@@ -54,7 +54,7 @@ You can either initialize by passing authentication or by setting an environment
    from langgraph_sdk import get_client

    # only pass the url argument to get_client() if you changed the default port when calling langgraph up
-    client = get_client(url=<DEPLOYMENT_URL>,api_key=<LANGCHAIN_API_KEY>)
+    client = get_client(url=<DEPLOYMENT_URL>,api_key=<LANGSMITH_API_KEY>)
    # Using the graph deployed with the name "agent"
    assistant_id = "agent"
    thread = await client.threads.create()
@@ -66,7 +66,7 @@ You can either initialize by passing authentication or by setting an environment
    import { Client } from "@langchain/langgraph-sdk";

    // only set the apiUrl if you changed the default port when calling langgraph up
-    const client = new Client({ apiUrl: <DEPLOYMENT_URL>, apiKey: <LANGCHAIN_API_KEY> });
+    const client = new Client({ apiUrl: <DEPLOYMENT_URL>, apiKey: <LANGSMITH_API_KEY> });
    // Using the graph deployed with the name "agent"
    const assistantId = "agent";
    const thread = await client.threads.create();
@@ -78,13 +78,13 @@ You can either initialize by passing authentication or by setting an environment
    curl --request POST \
      --url <DEPLOYMENT_URL>/threads \
      --header 'Content-Type: application/json'
-      --header 'x-api-key: <LANGCHAIN_API_KEY>'
+      --header 'x-api-key: <LANGSMITH_API_KEY>'
    ```
  

 #### Initialize with environment variables

-If you have a `LANGCHAIN_API_KEY` set in your environment, you do not need to explicitly pass authentication to the client
+If you have a `LANGSMITH_API_KEY` set in your environment, you do not need to explicitly pass authentication to the client

 === "Python"

@@ -154,7 +154,7 @@ Now we can invoke our graph to ensure it is working. Make sure to change the inp
    }
    ```

-  === "CURL"
+=== "CURL"

    ```bash
    curl --request POST \
@@ -83,7 +83,7 @@ We can now call `.get_schemas` to get schemas associated with this graph:
        assistant_id=assistant["assistant_id"]
    )
    # There are multiple types of schemas
-    # We can get the `config_schema` to look at the the configurable parameters
+    # We can get the `config_schema` to look at the configurable parameters
    print(schemas["config_schema"])
    ```

@@ -94,7 +94,7 @@ We can now call `.get_schemas` to get schemas associated with this graph:
      assistant["assistant_id"]
    );
    // There are multiple types of schemas
-    // We can get the `config_schema` to look at the the configurable parameters
+    // We can get the `config_schema` to look at the configurable parameters
    console.log(schemas.config_schema);
    ```

@@ -8,9 +8,9 @@ If you want to learn how to build an agent like this from scratch, take a look a

 This tutorial will use:

- Anthropic for the LLM - sign up and get an API key [here](https://console.anthropic.com/)
- Tavily for the search engine - sign up and get an API key [here](https://app.tavily.com/)
- LangSmith for hosting - sign up and get an API key [here](https://smith.langchain.com/)
+- Anthropic for the LLM - sign up and get an API key [here](https://console.anthropic.com/).
+- Tavily for the search engine - sign up and get an API key [here](https://app.tavily.com/).
+- LangSmith for hosting - sign up and get an API key [here](https://smith.langchain.com/).

 ## Create and configure your app

@@ -1,19 +0,0 @@
-<!doctype html>
-<html>
-  <head>
-    <title>Open Assistants API Specification</title>
-    <meta charset="utf-8" />
-    <meta
-      name="viewport"
-      content="width=device-width, initial-scale=1" />
-  </head>
-  <body>
-    <script id="api-reference" data-url="./open_agent_api.json"></script>
-    <script>
-      var configuration = {}
-      document.getElementById('api-reference').dataset.configuration =
-        JSON.stringify(configuration)
-    </script>
-    <script src="https://cdn.jsdelivr.net/npm/@scalar/api-reference"></script>
-  </body>
-</html>
@@ -1557,8 +1557,11 @@
          "200": {
            "description": "Success",
            "content": {
-              "application/json": {
-                "schema": {}
+              "text/event-stream": {
+                "schema": {
+                  "type": "string",
+                  "description": "The server will send a stream of events in SSE format.\n\n**Example event**:\n\nid: 1\n\nevent: message\n\ndata: {}"
+                }
              }
            }
          },
@@ -1905,8 +1908,11 @@
          "200": {
            "description": "Success",
            "content": {
-              "application/json": {
-                "schema": {}
+              "text/event-stream": {
+                "schema": {
+                  "type": "string",
+                  "description": "The server will send a stream of events in SSE format.\n\n**Example event**:\n\nid: 1\n\nevent: message\n\ndata: {}"
+                }
              }
            }
          },
@@ -2143,8 +2149,11 @@
          "200": {
            "description": "Success",
            "content": {
-              "application/json": {
-                "schema": {}
+              "text/event-stream": {
+                "schema": {
+                  "type": "string",
+                  "description": "The server will send a stream of events in SSE format.\n\n**Example event**:\n\nid: 1\n\nevent: message\n\ndata: {}"
+                }
              }
            }
          },
@@ -26,10 +26,11 @@ The LangGraph command line interface includes commands to build and run a LangGr
 The LangGraph CLI requires a JSON configuration file with the following keys:

 | Key                | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
-|--------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
 | `dependencies`     | **Required**. Array of dependencies for LangGraph Cloud API server. Dependencies can be one of the following: (1) `"."`, which will look for local Python packages, (2) `pyproject.toml`, `setup.py` or `requirements.txt` in the app directory `"./local_package"`, or (3) a package name.                                                                                                                                                                                                                                                  |
 | `graphs`           | **Required**. Mapping from graph ID to path where the compiled graph or a function that makes a graph is defined. Example: <ul><li>`./your_package/your_file.py:variable`, where `variable` is an instance of `langgraph.graph.state.CompiledStateGraph`</li><li>`./your_package/your_file.py:make_graph`, where `make_graph` is a function that takes a config dictionary (`langchain_core.runnables.RunnableConfig`) and creates an instance of `langgraph.graph.state.StateGraph` / `langgraph.graph.state.CompiledStateGraph`.</li></ul> |
 | `env`              | Path to `.env` file or a mapping from environment variable to its value.                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
+| `store`            | Configuration for adding semantic search to the BaseStore. Contains the following fields: <ul><li>`index`: Configuration for semantic search indexing with fields:<ul><li>`embed`: Embedding provider (e.g., "openai:text-embedding-3-small") or path to custom embedding function</li><li>`dims`: Dimension size of the embedding model. Used to initialize the vector table.</li><li>`fields` (optional): List of fields to index. Defaults to `["$"]`, meaningto index entire documents. Can be specific fields like `["text", "summary", "some.value"]`</li></ul></li></ul>                                                                  |
 | `python_version`   | `3.11` or `3.12`. Defaults to `3.11`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
 | `pip_config_file`  | Path to `pip` config file.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
 | `dockerfile_lines` | Array of additional lines to add to Dockerfile following the import from parent image.                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
@@ -41,33 +42,84 @@ The LangGraph CLI requires a JSON configuration file with the following keys:
    </p>
 </div>

-Example:
+### Examples
+
+#### Basic Configuration

 ```json
 {
-  "dependencies": ["langchain_openai", "./your_package"],
+  "dependencies": ["."],
  "graphs": {
-    "my_graph_id": "./your_package/your_file.py:variable"
-  },
-  "env": "./.env"
+    "chat": "./chat/graph.py:graph"
+  }
 }
 ```

-Example:
+#### Adding semantic search to the store
+
+All deployments come with a DB-backed BaseStore. Adding an "index" configuration to your `langgraph.json` will enable [semantic search](../deployment/semantic_search.md) within the BaseStore of your deployment.
+
+The `fields` configuration determines which parts of your documents to embed:
+- If omitted or set to `["$"]`, the entire document will be embedded
+- To embed specific fields, use JSON path notation: `["metadata.title", "content.text"]`
+- Documents missing specified fields will still be stored but won't have embeddings for those fields
+- You can still override which fields to embed on a specific item at `put` time using the `index` parameter

 ```json
 {
-  "python_version": "3.11",
-  "dependencies": ["langchain_openai", "."],
+  "dependencies": ["."],
  "graphs": {
-    "my_graph_id": "./your_package/your_file.py:make_graph"
+    "memory_agent": "./agent/graph.py:graph"
  },
-  "env": {
-    "OPENAI_API_KEY": "secret-key"
+  "store": {
+    "index": {
+      "embed": "openai:text-embedding-3-small",
+      "dims": 1536,
+      "fields": ["$"]
+    }
  }
 }
 ```

+!!! note "Common model dimensions"
+        - openai:text-embedding-3-large: 3072
+        - openai:text-embedding-3-small: 1536
+        - openai:text-embedding-ada-002: 1536
+        - cohere:embed-english-v3.0: 1024
+        - cohere:embed-english-light-v3.0: 384
+        - cohere:embed-multilingual-v3.0: 1024
+        - cohere:embed-multilingual-light-v3.0: 384
+
+#### Semantic search with a custom embedding function
+
+If you want to use semantic search with a custom embedding function, you can pass a path to a custom embedding function:
+
+```json
+{
+  "dependencies": ["."],
+  "graphs": {
+    "memory_agent": "./agent/graph.py:graph"
+  },
+  "store": {
+    "index": {
+      "embed": "./embeddings.py:embed_texts",
+      "dims": 768,
+      "fields": ["text", "summary"]
+    }
+  }
+}
+```
+
+The `embed` field in store configuration can reference a custom function that takes a list of strings and returns a list of embeddings. Example implementation:
+
+```python
+# embeddings.py
+def embed_texts(texts: list[str]) -> list[list[float]]:
+    """Custom embedding function for semantic search."""
+    # Implementation using your preferred embedding model
+    return [[0.1, 0.2, ...] for _ in texts]  # dims-dimensional vectors
+```
+
 ## Commands

 The base command for the LangGraph CLI is `langgraph`.
@@ -78,6 +130,42 @@ The base command for the LangGraph CLI is `langgraph`.
 langgraph [OPTIONS] COMMAND [ARGS]
 ```

+### `dev`
+
+Run LangGraph API server in development mode with hot reloading and debugging capabilities. This lightweight server requires no Docker installation and is suitable for development and testing. State is persisted to a local directory.
+
+!!! note "Python only"
+
+    Currently, the CLI only supports Python >= 3.11.
+    JS support is coming soon.
+
+**Installation**
+
+This command requires the "inmem" extra to be installed:
+
+```bash
+pip install -U "langgraph-cli[inmem]"
+```
+
+**Usage**
+
+```
+langgraph dev [OPTIONS]
+```
+
+**Options**
+
+| Option                        | Default          | Description                                                                         |
+| ----------------------------- | ---------------- | ----------------------------------------------------------------------------------- |
+| `-c, --config FILE`           | `langgraph.json` | Path to configuration file declaring dependencies, graphs and environment variables |
+| `--host TEXT`                 | `127.0.0.1`      | Host to bind the server to                                                          |
+| `--port INTEGER`              | `2024`           | Port to bind the server to                                                          |
+| `--no-reload`                 |                  | Disable auto-reload                                                                 |
+| `--n-jobs-per-worker INTEGER` |                  | Number of jobs per worker. Default is 10                                            |
+| `--no-browser`                |                  | Disable automatic browser opening                                                   |
+| `--debug-port INTEGER`        |                  | Port for debugger to listen on                                                      |
+| `--help`                      |                  | Display command documentation                                                       |
+
 ### `build`

 Build LangGraph Cloud API server Docker image.
@@ -91,7 +179,7 @@ langgraph build [OPTIONS]
 **Options**

 | Option               | Default          | Description                                                                                                                  |
-|----------------------|------------------|------------------------------------------------------------------------------------------------------------------------------|
+| -------------------- | ---------------- | ---------------------------------------------------------------------------------------------------------------------------- |
 | `--platform TEXT`    |                  | Target platform(s) to build the Docker image for. Example: `langgraph build --platform linux/amd64,linux/arm64`              |
 | `-t, --tag TEXT`     |                  | **Required**. Tag for the Docker image. Example: `langgraph build -t my-image`                                               |
 | `--pull / --no-pull` | `--pull`         | Build with latest remote Docker image. Use `--no-pull` for running the LangGraph Cloud API server with locally built images. |
@@ -100,7 +188,7 @@ langgraph build [OPTIONS]

 ### `up`

-Start langgraph API server. For local testing, requires a LangSmith API key with access to LangGraph Cloud closed beta. Requires a license key for production use.
+Start LangGraph API server. For local testing, requires a LangSmith API key with access to LangGraph Cloud closed beta. Requires a license key for production use.

 **Usage**

@@ -110,20 +198,20 @@ langgraph up [OPTIONS]

 **Options**

-| Option                       | Default                   | Description                                                                                                           |
-|------------------------------|---------------------------|-----------------------------------------------------------------------------------------------------------------------|
-| `--wait`                     |                           | Wait for services to start before returning. Implies --detach                                                         |
-| `--postgres-uri TEXT`        | Local database            | Postgres URI to use for the database.                                                                                 |
-| `--watch`                    |                           | Restart on file changes                                                                                               |
-| `--debugger-base-url TEXT`   | `http://127.0.0.1:[PORT]` | URL used by the debugger to access LangGraph API.                                                                     |
-| `--debugger-port INTEGER`    |                           | Pull the debugger image locally and serve the UI on specified port                                                    |
-| `--verbose`                  |                           | Show more output from the server logs.                                                                                |
-| `-c, --config FILE`          | `langgraph.json`          | Path to configuration file declaring dependencies, graphs and environment variables.                                  |
-| `-d, --docker-compose FILE`  |                           | Path to docker-compose.yml file with additional services to launch.                                                   |
-| `-p, --port INTEGER`         | `8123`                    | Port to expose. Example: `langgraph test --port 8000`                                                                 |
-| `--pull / --no-pull`         | `pull`                    | Pull latest images. Use --no-pull for running the server with locally-built images. Example: `langgraph up --no-pull` |
-| `--recreate / --no-recreate` | `no-recreate`             | Recreate containers even if their configuration and image haven't changed                                             |
-| `--help`                     |                           | Display command documentation.                                                                                        |
+| Option                       | Default                   | Description                                                                                                             |
+| ---------------------------- | ------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
+| `--wait`                     |                           | Wait for services to start before returning. Implies --detach                                                           |
+| `--postgres-uri TEXT`        | Local database            | Postgres URI to use for the database.                                                                                   |
+| `--watch`                    |                           | Restart on file changes                                                                                                 |
+| `--debugger-base-url TEXT`   | `http://127.0.0.1:[PORT]` | URL used by the debugger to access LangGraph API.                                                                       |
+| `--debugger-port INTEGER`    |                           | Pull the debugger image locally and serve the UI on specified port                                                      |
+| `--verbose`                  |                           | Show more output from the server logs.                                                                                  |
+| `-c, --config FILE`          | `langgraph.json`          | Path to configuration file declaring dependencies, graphs and environment variables.                                    |
+| `-d, --docker-compose FILE`  |                           | Path to docker-compose.yml file with additional services to launch.                                                     |
+| `-p, --port INTEGER`         | `8123`                    | Port to expose. Example: `langgraph up --port 8000`                                                                     |
+| `--pull / --no-pull`         | `pull`                    | Pull latest images. Use `--no-pull` for running the server with locally-built images. Example: `langgraph up --no-pull` |
+| `--recreate / --no-recreate` | `no-recreate`             | Recreate containers even if their configuration and image haven't changed                                               |
+| `--help`                     |                           | Display command documentation.                                                                                          |

 ### `dockerfile`

@@ -138,7 +226,7 @@ langgraph dockerfile [OPTIONS] SAVE_PATH
 **Options**

 | Option              | Default          | Description                                                                                                     |
-|---------------------|------------------|-----------------------------------------------------------------------------------------------------------------|
+| ------------------- | ---------------- | --------------------------------------------------------------------------------------------------------------- |
 | `-c, --config FILE` | `langgraph.json` | Path to the [configuration file](#configuration-file) declaring dependencies, graphs and environment variables. |
 | `--help`            |                  | Show this message and exit.                                                                                     |

@@ -148,9 +236,9 @@ Example:
 langgraph dockerfile -c langgraph.json Dockerfile
 ```

-Would generate something like the following:
+This generates a Dockerfile that looks similar to:

-```text
+```dockerfile
 FROM langchain/langgraph-api:3.11

 ADD ./pipconf.txt /pipconfig.txt
@@ -170,6 +258,3 @@ RUN set -ex && \
 RUN PIP_CONFIG_FILE=/pipconfig.txt PYTHONDONTWRITEBYTECODE=1 pip install --no-cache-dir -c /api/constraints.txt -e /deps/*

 ENV LANGSERVE_GRAPHS='{"agent": "/deps/__outer_graphs/src/agent.py:graph", "storm": "/deps/__outer_graphs/src/storm.py:graph"}'
-```
-
-You can then customize, build images, push, and deploy from this file.
@@ -2,13 +2,13 @@

 LLMs are extremely powerful, particularly when connected to other systems such as a retriever or APIs. This is why many LLM applications use a control flow of steps before and / or after LLM calls. As an example [RAG](https://github.com/langchain-ai/rag-from-scratch) performs retrieval of relevant documents to a question, and passes those documents to an LLM in order to ground the response. Often a control flow of steps before and / or after an LLM is called a "chain." Chains are a popular paradigm for programming with LLMs and offer a high degree of reliability; the same set of steps runs with each chain invocation.

-However, we often want LLM systems that can pick their own control flow! This is one definition of an [agent](https://blog.langchain.dev/what-is-an-agent/): an agent is a system that uses an LLM to decide the control flow of an application. Unlike a chain, an agent given an LLM some degree of control over the sequence of steps in the application. Examples of using an LLM to decide the control of an application:
+However, we often want LLM systems that can pick their own control flow! This is one definition of an [agent](https://blog.langchain.dev/what-is-an-agent/): an agent is a system that uses an LLM to decide the control flow of an application. Unlike a chain, an agent gives an LLM some degree of control over the sequence of steps in the application. Examples of using an LLM to decide the control of an application:

 - Using an LLM to route between two potential paths
 - Using an LLM to decide which of many tools to call
 - Using an LLM to decide whether the generated answer is sufficient or more work is need

-There are many different types of [agent architectures](https://blog.langchain.dev/what-is-a-cognitive-architecture/) to consider, which given an LLM varying levels of control. On one extreme, a router allows an LLM to select a single step from a specified set of options and, on the other extreme, a fully autonomous long-running agent may have complete freedom to select any sequence of steps that it wants for a given problem. 
+There are many different types of [agent architectures](https://blog.langchain.dev/what-is-a-cognitive-architecture/) to consider, which give an LLM varying levels of control. On one extreme, a router allows an LLM to select a single step from a specified set of options and, on the other extreme, a fully autonomous long-running agent may have complete freedom to select any sequence of steps that it wants for a given problem. 

 ![Agent Types](img/agent_types.png)

@@ -27,8 +27,8 @@ Adding a [breakpoint](./low_level.md#breakpoints) a specific location in the gra
 Here, we compile our graph with a checkpointer and a breakpoint at the node we want to interrupt before, `step_for_human_in_the_loop`. We then perform one of the above interaction patterns, which will create a new checkpoint if a human edits the graph state. The new checkpoint is saved to the `thread` and we can resume the graph execution from there by passing in `None` as the input.

 ```python
-# Compile our graph with a checkpoitner and a breakpoint before "step_for_human_in_the_loop"
-graph = builder.compile(checkpointer=checkpoitner, interrupt_before=["step_for_human_in_the_loop"])
+# Compile our graph with a checkpointer and a breakpoint before "step_for_human_in_the_loop"
+graph = builder.compile(checkpointer=checkpointer, interrupt_before=["step_for_human_in_the_loop"])

 # Run the graph up to the breakpoint
 thread_config = {"configurable": {"thread_id": "1"}}
@@ -98,8 +98,8 @@ With persistence, we can surface the current agent state as well as the next ste
 If approved, the graph resumes execution from the last saved checkpoint, which is saved to the `thread`:

 ```python
-# Compile our graph with a checkpoitner and a breakpoint before the step to approve
-graph = builder.compile(checkpointer=checkpoitner, interrupt_before=["node_2"])
+# Compile our graph with a checkpointer and a breakpoint before the step to approve
+graph = builder.compile(checkpointer=checkpointer, interrupt_before=["node_2"])

 # Run the graph up to the breakpoint
 for event in graph.stream(inputs, thread, stream_mode="values"):
@@ -120,7 +120,7 @@ See [our guide](../how-tos/human_in_the_loop/breakpoints.ipynb) for a detailed h

 Sometimes we want to review and edit the agent's state. 
 
-As with approval, we can interrupt our agent at a [breakpoint](./low_level.md#breakpoints) prior the the step we want to check. 
+As with approval, we can interrupt our agent at a [breakpoint](./low_level.md#breakpoints) prior to the step we want to check. 
 
 We can surface the current state to a user and allow the user to edit the agent state.
 
@@ -131,8 +131,8 @@ We can edit the graph state by forking the current checkpoint, which is saved to
 We can then proceed with the graph from our forked checkpoint as done before. 

 ```python
-# Compile our graph with a checkpoitner and a breakpoint before the step to review
-graph = builder.compile(checkpointer=checkpoitner, interrupt_before=["node_2"])
+# Compile our graph with a checkpointer and a breakpoint before the step to review
+graph = builder.compile(checkpointer=checkpointer, interrupt_before=["node_2"])

 # Run the graph up to the breakpoint
 for event in graph.stream(inputs, thread, stream_mode="values"):
@@ -170,11 +170,11 @@ With editing, the user makes a decision about whether or not to edit the graph s

 With input, we explicitly define a node in our graph for collecting human input!

-The the state update with the human input then runs *as this node*.
+The state update with the human input then runs *as this node*.

 ```python
-# Compile our graph with a checkpoitner and a breakpoint before the step to to collect human input
-graph = builder.compile(checkpointer=checkpoitner, interrupt_before=["human_input"])
+# Compile our graph with a checkpointer and a breakpoint before the step to to collect human input
+graph = builder.compile(checkpointer=checkpointer, interrupt_before=["human_input"])

 # Run the graph up to the breakpoint
 for event in graph.stream(inputs, thread, stream_mode="values"):
@@ -211,8 +211,8 @@ Even if the tool call is correct, we may also want to apply discretion:
 With these points in mind, we can combine the above ideas to create a human-in-the-loop review of a tool call.

 ```python
-# Compile our graph with a checkpoitner and a breakpoint before the step to to review the tool call from the LLM 
-graph = builder.compile(checkpointer=checkpoitner, interrupt_before=["human_review"])
+# Compile our graph with a checkpointer and a breakpoint before the step to to review the tool call from the LLM 
+graph = builder.compile(checkpointer=checkpointer, interrupt_before=["human_review"])

 # Run the graph up to the breakpoint
 for event in graph.stream(inputs, thread, stream_mode="values"):
@@ -319,4 +319,4 @@ for event in graph.stream(None, config, stream_mode="values"):

 See [this additional conceptual guide](https://langchain-ai.github.io/langgraph/concepts/persistence/#update-state) for related context on forking.

-See see [this guide](../how-tos/human_in_the_loop/time-travel.ipynb) for a detailed how-to on doing time-travel!
+See see [this guide](../how-tos/human_in_the_loop/time-travel.ipynb) for a detailed how-to on doing time-travel!
@@ -30,7 +30,7 @@ The conceptual guide does not cover step-by-step instructions or specific implem
 - [Streaming](streaming.md): Streaming is crucial for enhancing the responsiveness of applications built on LLMs. By displaying output progressively, even before a complete response is ready, streaming significantly improves user experience (UX), particularly when dealing with the latency of LLMs. 
 - [FAQ](faq.md): Frequently asked questions about LangGraph.

-## LangGraph Platform 
+## LangGraph Platform

 LangGraph Platform is a commercial solution for deploying agentic applications in production, built on the open-source LangGraph framework.

@@ -28,9 +28,33 @@ The CLI provides the following core functionality:

 The `langgraph build` command builds a Docker image for the [LangGraph API server](./langgraph_server.md) that can be directly deployed.

+### `dev`
+
+!!! note "New in version 0.1.55"
+    The `langgraph dev` command was introduced in langgraph-cli version 0.1.55.
+
+!!! note "Python only"
+
+    Currently, the CLI only supports Python >= 3.11.
+    JS support is coming soon.
+
+The `langgraph dev` command starts a lightweight development server that requires no Docker installation. This server is ideal for rapid development and testing, with features like:
+
+- Hot reloading: Changes to your code are automatically detected and reloaded
+- Debugger support: Attach your IDE's debugger for line-by-line debugging
+- In-memory state with local persistence: Server state is stored in memory for speed but persisted locally between restarts
+
+To use this command, you need to install the CLI with the "inmem" extra:
+
+```bash
+pip install -U "langgraph-cli[inmem]"
+```
+
+**Note**: This command is intended for local development and testing only. It is not recommended for production use. Since it does not use Docker, we recommend using virtual environments to manage your project's dependencies.
+
 ### `up`

-The `langgraph up` command starts an instance of the [LangGraph API server](./langgraph_server.md) locally. This requires docker to be installed and running locally. It also requires a LangSmith API key for local development or a license key for production use.
+The `langgraph up` command starts an instance of the [LangGraph API server](./langgraph_server.md) locally in a docker container. This requires thedocker server to be running locally. It also requires a LangSmith API key for local development or a license key for production use.

 The server includes all API endpoints for your graph's runs, threads, assistants, etc. as well as the other services required to run your agent, including a managed database for checkpointing and storage.

@@ -14,6 +14,13 @@ A **deployment** is an instance of a LangGraph API. A single deployment can have

 See the [how-to guide](../cloud/deployment/cloud.md#create-new-deployment) for creating a new deployment.

+## Resource Allocation
+
+| **Deployment Type** | **CPU** | **Memory** | **Scaling**         |
+|---------------------|---------|------------|---------------------|
+| Development         | 1 CPU   | 1 GB       | Up to 1 container   |
+| Production          | 1 CPU   | 2 GB       | Up to 10 containers |
+
 ## Revision

 A revision is an iteration of a [deployment](#deployment). When a new deployment is created, an initial revision is automatically created. To deploy new code changes or update environment variable configurations for a deployment, a new revision must be created. When a revision is created, a new container image is built automatically.
@@ -33,6 +40,7 @@ A high-level diagram of a Cloud SaaS deployment.

 ![diagram](img/langgraph_cloud_architecture.png)

+
 ## Related

 - [Deployment Options](./deployment_options.md)
@@ -35,6 +35,16 @@ While in Beta, LangGraph Studio is available for free to all [LangSmith](https:/

 If you have deployed your LangGraph application on LangGraph Platform (Cloud), you can access the studio as part of that

+### Development server
+
+LangGraph CLI also contains a command for running an in-memory development server that can be used to connect a local LangGraph app with the studio.
+See [instructions here](../cloud/reference/cli.md#dev) for more information.
+
+The way this works is that it runs inside your local environment.
+It will spin up an in-memory, development server to deploy the graph.
+You can then connect to the studio via the Cloud hosted version of LangGraph Platform.
+To be clear, the web studio will connect to your locally running server - your agent is still running locally and never leaves your device.
+
 ## Studio FAQs

 ### Why is my project failing to start?
@@ -283,6 +283,9 @@ You can optionally provide a dictionary that maps the `routing_function`'s outpu
 graph.add_conditional_edges("node_a", routing_function, {True: "node_b", False: "node_c"})
 ```

+!!! tip
+    Use [`Command`](#command) instead of conditional edges if you want to combine state updates and routing in a single function.
+
 ### Entry Point

 The entry point is the first node(s) that are run when the graph starts. You can use the [`add_edge`][langgraph.graph.StateGraph.add_edge] method from the virtual [`START`][langgraph.constants.START] node to the first node to execute to specify where to enter the graph.
@@ -322,6 +325,64 @@ def continue_to_jokes(state: OverallState):
 graph.add_conditional_edges("node_a", continue_to_jokes)
 ```

+## `Command`
+
+It can be useful to combine control flow (edges) and state updates (nodes). For example, you might want to BOTH perform state updates AND decide which node to go to next in the SAME node. LangGraph provides a way to do so by returning a [`Command`][langgraph.types.Command] object from node functions:
+
+```python
+def my_node(state: State) -> Command[Literal["my_other_node"]]:
+    return Command(
+        # state update
+        update={"foo": "bar"},
+        # control flow
+        goto="my_other_node"
+    )
+```
+
+With `Command` you can also achieve dynamic control flow behavior (identical to [conditional edges](#conditional-edges)):
+
+```python
+def my_node(state: State) -> Command[Literal["my_other_node"]]:
+    if state["foo"] == "bar":
+        return Command(update={"foo": "baz"}, goto="my_other_node")
+```
+
+!!! important
+
+    When returning `Command` in your node functions, you must add return type annotations with the list of node names the node is routing to, e.g. `Command[Literal["my_other_node"]]`. This is necessary for the graph rendering and tells LangGraph that `my_node` can navigate to `my_other_node`.
+
+Check out this [how-to guide](../how-tos/command.ipynb) for an end-to-end example of how to use `Command`.
+
+### When should I use Command instead of conditional edges?
+
+Use `Command` when you need to **both** update the graph state **and** route to a different node. For example, when implementing [multi-agent handoffs](./multi_agent.md#handoffs) where it's important to route to a different agent and pass some information to that agent.
+
+Use [conditional edges](#conditional-edges) to route between nodes conditionally without updating the state.
+
+### Using inside tools
+
+A common use case is updating graph state from inside a tool. For example, in a customer support application you might want to look up customer information based on their account number or ID in the beginning of the conversation. To update the graph state from the tool, you can return `Command(update={"my_custom_key": "foo", "messages": [...]})` from the tool:
+
+```python
+@tool
+def lookup_user_info(tool_call_id: Annotated[str, InjectedToolCallId], config: RunnableConfig):
+    """Use this to look up user information to better assist them with their questions."""
+    user_info = get_user_info(config.get("configurable", {}).get("user_id"))
+    return Command(
+        update={
+            # update the state keys
+            "user_info": user_info,
+            # update the message history
+            "messages": [ToolMessage("Successfully looked up user information", tool_call_id=tool_call_id)]
+        }
+    )
+```
+
+!!! important
+    You MUST include `messages` (or any state key used for the message history) in `Command.update` when returning `Command` from a tool and the list of messages in `messages` MUST contain a `ToolMessage`. This is necessary for the resulting message history to be valid (LLM providers require AI messages with tool calls to be followed by the tool result messages).
+
+If you are using tools that update state via `Command`, we recommend using prebuilt [`ToolNode`][langgraph.prebuilt.tool_node.ToolNode] which automatically handles tools returning `Command` objects and propagates them to the graph state. If you're writing a custom node that calls tools, you would need to manually propagate `Command` objects returned by the tools as the update from node.
+
 ## Persistence

 LangGraph provides built-in persistence for your agent's state using [checkpointers][langgraph.checkpoint.base.BaseCheckpointSaver]. Checkpointers save snapshots of the graph state at every superstep, allowing resumption at any time. This enables features like human-in-the-loop interactions, memory management, and fault-tolerance. You can even directly manipulate a graph's state after its execution using the
@@ -171,7 +171,7 @@ trim_messages(

 ## Long-term memory

-Long-term memory in LangGraph allows systems to retain information across different conversations or sessions. Unlike short-term memory, which is thread-scoped, long-term memory is saved within custom "namespaces."
+Long-term memory in LangGraph allows systems to retain information across different conversations or sessions. Unlike short-term memory, which is **thread-scoped**, long-term memory is saved within custom "namespaces."

 ### Storing memories

@@ -180,16 +180,34 @@ LangGraph stores long-term memories as JSON documents in a [store](persistence.m
 ```python
 from langgraph.store.memory import InMemoryStore

+
+def embed(texts: list[str]) -> list[list[float]]:
+    # Replace with an actual embedding function or LangChain embeddings object
+    return [[1.0, 2.0] * len(texts)]
+
+
 # InMemoryStore saves data to an in-memory dictionary. Use a DB-backed store in production use.
-store = InMemoryStore()
+store = InMemoryStore(index={"embed": embed, "dims": 2})
 user_id = "my-user"
 application_context = "chitchat"
 namespace = (user_id, application_context)
-store.put(namespace, "a-memory", {"rules": ["User likes short, direct language", "User only speaks English & python"], "my-key": "my-value"})
+store.put(
+    namespace,
+    "a-memory",
+    {
+        "rules": [
+            "User likes short, direct language",
+            "User only speaks English & python",
+        ],
+        "my-key": "my-value",
+    },
+)
 # get the "memory" by ID
 item = store.get(namespace, "a-memory")
-# list "memories" within this namespace, filtering on content equivalence
-items = store.search(namespace, filter={"my-key": "my-value"})
+# search for "memories" within this namespace, filtering on content equivalence, sorted by vector similarity
+items = store.search(
+    namespace, filter={"my-key": "my-value"}, query="language preferences"
+)
 ```

 ### Framework for thinking about long-term memory
@@ -218,6 +236,9 @@ Different applications require various types of memory. Although the analogy isn

 [Semantic memory](https://en.wikipedia.org/wiki/Semantic_memory), both in humans and AI agents, involves the retention of specific facts and concepts. In humans, it can include information learned in school and the understanding of concepts and their relationships. For AI agents, semantic memory is often used to personalize applications by remembering facts or concepts from past interactions. 

+> Note: Not to be confused with "semantic search" which is a technique for finding similar content using "meaning" (usually as embeddings). Semantic memory is a term from psychology, referring to storing facts and knowledge, while semantic search is a method for retrieving information based on meaning rather than exact matches.
+
+
 #### Profile

 Semantic memories can be managed in different ways. For example, memories can be a single, continuously updated "profile" of well-scoped and specific information about a user, organization, or other entity (including the agent itself). A profile is generally just a JSON document with various key-value pairs you've selected to represent your domain. 
@@ -232,7 +253,7 @@ Alternatively, memories can be a collection of documents that are continuously u

 However, this shifts some complexity memory updating. The model must now _delete_ or _update_ existing items in the list, which can be tricky. In addition, some models may default to over-inserting and others may default to over-updating. See the [Trustcall](https://github.com/hinthornw/trustcall) package for one way to manage this and consider evaluation (e.g., with a tool like [LangSmith](https://docs.smith.langchain.com/tutorials/Developers/evaluation)) to help you tune the behavior.

-Working with document collections also shifts complexity to memory **search** over the list. The `Store` currently supports [filtering by metadata](https://langchain-ai.github.io/langgraph/reference/store/#storage) and will soon add [semantic search shortly](https://python.langchain.com/docs/concepts/vectorstores/), but selecting the most relevant documents can be tricky as the list grows.
+Working with document collections also shifts complexity to memory **search** over the list. The `Store` currently supports both [semantic search](https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.SearchOp.query) and [filtering by content](https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.SearchOp.filter).

 Finally, using a collection of memories can make it challenging to provide comprehensive context to the model. While individual memories may follow a specific schema, this structure might not capture the full context or relationships between memories. As a result, when using these memories to generate responses, the model may lack important contextual information that would be more readily available in a unified profile approach.

@@ -26,18 +26,88 @@ There are several ways to connect agents in a multi-agent system:
 - **Hierarchical**: you can define a multi-agent system with [a supervisor of supervisors](https://langchain-ai.github.io/langgraph/tutorials/multi_agent/hierarchical_agent_teams/). This is a generalization of the supervisor architecture and allows for more complex control flows.
 - **Custom multi-agent workflow**: each agent communicates with only a subset of agents. Parts of the flow are deterministic, and only some agents can decide which other agents to call next.

+### Handoffs
+
+In multi-agent architectures, agents can be represented as graph nodes. Each agent node executes its step(s) and decides whether to finish execution or route to another agent, including potentially routing to itself (e.g., running in a loop). A common pattern in multi-agent interactions is handoffs, where one agent hands off control to another. Handoffs allow you to specify:
+
+- __destination__: target agent to navigate to (e.g., name of the node to go to)
+- __payload__: [information to pass to that agent](#communication-between-agents) (e.g., state update)
+
+To implement handoffs in LangGraph, agent nodes can return [`Command`](./low_level.md#command) object that allows you to combine both control flow and state updates:
+
+```python
+def agent(state) -> Command[Literal["agent", "another_agent"]]:
+    # the condition for routing/halting can be anything, e.g. LLM tool call / structured output, etc.
+    goto = get_next_agent(...)  # 'agent' / 'another_agent'
+    return Command(
+        # Specify which agent to call next
+        goto=goto,
+        # Update the graph state
+        update={"my_state_key": "my_state_value"}
+    )
+```
+
+In a more complex scenario where each agent node is itself a graph (i.e., a [subgraph](./low_level.md#subgraphs)), a node in one of the agent subgraphs might want to navigate to a different agent. For example, if you have two agents, `alice` and `bob` (subgraph nodes in a parent graph), and `alice` needs to navigate to `bob`, you can set `graph=Command.PARENT` in the `Command` object:
+
+```python
+def some_node_inside_alice(state)
+    return Command(
+        goto="bob",
+        update={"my_state_key": "my_state_value"},
+        # specify which graph to navigate to (defaults to the current graph)
+        graph=Command.PARENT,
+    )
+```
+
+!!! note
+    If you need to support visualization for subgraphs communicating using `Command(graph=Command.PARENT)` you would need to wrap them in a node function with `Command` annotation, e.g. instead of this:
+
+    ```python
+    builder.add_node(alice)
+    ```
+
+    you would need to do this:
+
+    ```python
+    def call_alice(state) -> Command[Literal["bob"]]:
+        return alice.invoke(state)
+
+    builder.add_node("alice", call_alice)
+    ```
+
+#### Handoffs as tools
+
+One of the most common agent types is a ReAct-style tool-calling agents. For those types of agents, a common pattern is wrapping a handoff in a tool call, e.g.:
+
+```python
+def transfer_to_bob(state):
+    """Transfer to bob."""
+    return Command(
+        goto="bob",
+        update={"my_state_key": "my_state_value"},
+        graph=Command.PARENT,
+    )
+```
+
+This is a special case of updating the graph state from tools where in addition the state update, the control flow is included as well.
+
+!!! important
+
+    If you want to use tools that return `Command`, you can either use prebuilt [`create_react_agent`][langgraph.prebuilt.chat_agent_executor.create_react_agent] / [`ToolNode`][langgraph.prebuilt.tool_node.ToolNode] components, or implement your own tool-executing node that collects `Command` objects returned by the tools and returns a list of them, e.g.:
+
+```python
+def call_tools(state):
+    ...
+    commands = [tools_by_name[call["name"].invoke(call, config={"coerce_tool_content": False}) for tool_call in tool_calls]
+    return commands
+```
+
+Let's now take a closer look at the different multi-agent architectures.
+
 ### Network

-In this architecture, agents are defined as graph nodes. Each agent can communicate with every other agent (many-to-many connections) and can decide which agent to call next. While very flexible, this architecture doesn't scale well as the number of agents grows:
+In this architecture, agents are defined as graph nodes. Each agent can communicate with every other agent (many-to-many connections) and can decide which agent to call next. This architecture is good for problems that do not have a clear hierarchy of agents or a specific sequence in which agents should be called.

- hard to enforce which agent should be called next
- hard to determine how much [information](#shared-message-list) should be passed between the agents
-
-We recommend avoiding this architecture in production and using one of the below architectures instead.
-
-### Supervisor
-
-In this architecture, we define agents as nodes and add a supervisor node (LLM) that decides which agent nodes should be called next. We use [conditional edges](./low_level.md#conditional-edges) to route execution to the appropriate agent node based on supervisor's decision. This architecture also lends itself well to running multiple agents in parallel or using [map-reduce](../how-tos/map-reduce.ipynb) pattern.

 ```python
 from typing import Literal
@@ -46,39 +116,83 @@ from langgraph.graph import StateGraph, MessagesState, START

 model = ChatOpenAI()

-class AgentState(MessagesState):
-    next: Literal["agent_1", "agent_2", "__end__"]
-
-def supervisor(state: AgentState):
+def agent_1(state: MessagesState) -> Command[Literal["agent_2", "agent_3", END]]:
    # you can pass relevant parts of the state to the LLM (e.g., state["messages"])
    # to determine which agent to call next. a common pattern is to call the model
    # with a structured output (e.g. force it to return an output with a "next_agent" field)
    response = model.invoke(...)
-    # the "next" key will be used by the conditional edges to route execution
-    # to the appropriate agent
-    return {"next": response["next_agent"]}
+    # route to one of the agents or exit based on the LLM's decision
+    # if the LLM returns "__end__", the graph will finish execution
+    return Command(
+        goto=response["next_agent"],
+        update={"messages": [response["content"]]},
+    )

-def agent_1(state: AgentState):
+def agent_2(state: MessagesState) -> Command[Literal["agent_1", "agent_3", END]]:
+    response = model.invoke(...)
+    return Command(
+        goto=response["next_agent"],
+        update={"messages": [response["content"]]},
+    )
+
+def agent_3(state: MessagesState) -> Command[Literal["agent_1", "agent_2", END]]:
+    ...
+    return Command(
+        goto=response["next_agent"],
+        update={"messages": [response["content"]]},
+    )
+
+builder = StateGraph(MessagesState)
+builder.add_node(agent_1)
+builder.add_node(agent_2)
+builder.add_node(agent_3)
+
+builder.add_edge(START, "agent_1")
+network = builder.compile()
+```
+
+### Supervisor
+
+In this architecture, we define agents as nodes and add a supervisor node (LLM) that decides which agent nodes should be called next. We use [`Command`](./low_level.md#command) to route execution to the appropriate agent node based on supervisor's decision. This architecture also lends itself well to running multiple agents in parallel or using [map-reduce](../how-tos/map-reduce.ipynb) pattern.
+
+```python
+from typing import Literal
+from langchain_openai import ChatOpenAI
+from langgraph.graph import StateGraph, MessagesState, START, END
+
+model = ChatOpenAI()
+
+def supervisor(state: MessagesState) -> Command[Literal["agent_1", "agent_2", END]]:
+    # you can pass relevant parts of the state to the LLM (e.g., state["messages"])
+    # to determine which agent to call next. a common pattern is to call the model
+    # with a structured output (e.g. force it to return an output with a "next_agent" field)
+    response = model.invoke(...)
+    # route to one of the agents or exit based on the supervisor's decision
+    # if the supervisor returns "__end__", the graph will finish execution
+    return Command(goto=response["next_agent"])
+
+def agent_1(state: MessagesState) -> Command[Literal["supervisor"]]:
    # you can pass relevant parts of the state to the LLM (e.g., state["messages"])
    # and add any additional logic (different models, custom prompts, structured output, etc.)
    response = model.invoke(...)
-    return {"messages": [response]}
+    return Command(
+        goto="supervisor",
+        update={"messages": [response]},
+    )

-def agent_2(state: AgentState):
+def agent_2(state: MessagesState) -> Command[Literal["supervisor"]]:
    response = model.invoke(...)
-    return {"messages": [response]}
+    return Command(
+        goto="supervisor",
+        update={"messages": [response]},
+    )

-builder = StateGraph(AgentState)
+builder = StateGraph(MessagesState)
 builder.add_node(supervisor)
 builder.add_node(agent_1)
 builder.add_node(agent_2)

 builder.add_edge(START, "supervisor")
-# route to one of the agents or exit based on the supervisor's decisiion
-# if the supervisor returns "__end__", the graph will finish execution
-builder.add_conditional_edges("supervisor", lambda state: state["next"])
-builder.add_edge("agent_1", "supervisor")
-builder.add_edge("agent_2", "supervisor")

 supervisor = builder.compile()
 ```
@@ -126,37 +240,29 @@ To address this, you can design your system _hierarchically_. For example, you c
 ```python
 from typing import Literal
 from langchain_openai import ChatOpenAI
-from langgraph.graph import StateGraph, MessagesState, START
+from langgraph.graph import StateGraph, MessagesState, START, END

 model = ChatOpenAI()

 # define team 1 (same as the single supervisor example above)
-class Team1State(MessagesState):
-    next: Literal["team_1_agent_1", "team_1_agent_2", "__end__"]

-def team_1_supervisor(state: Team1State):
+def team_1_supervisor(state: MessagesState) -> Command[Literal["team_1_agent_1", "team_1_agent_2", END]]:
    response = model.invoke(...)
-    return {"next": response["next_agent"]}
+    return Command(goto=response["next_agent"])

-def team_1_agent_1(state: Team1State):
+def team_1_agent_1(state: MessagesState) -> Command[Literal["team_1_supervisor"]]:
    response = model.invoke(...)
-    return {"messages": [response]}
+    return Command(goto="team_1_supervisor", update={"messages": [response]})

-def team_1_agent_2(state: Team1State):
+def team_1_agent_2(state: MessagesState) -> Command[Literal["team_1_supervisor"]]:
    response = model.invoke(...)
-    return {"messages": [response]}
+    return Command(goto="team_1_supervisor", update={"messages": [response]})

 team_1_builder = StateGraph(Team1State)
 team_1_builder.add_node(team_1_supervisor)
 team_1_builder.add_node(team_1_agent_1)
 team_1_builder.add_node(team_1_agent_2)
 team_1_builder.add_edge(START, "team_1_supervisor")
-# route to one of the agents or exit based on the supervisor's decisiion
-# if the supervisor returns "__end__", the graph will finish execution
-team_1_builder.add_conditional_edges("team_1_supervisor", lambda state: state["next"])
-team_1_builder.add_edge("team_1_agent_1", "team_1_supervisor")
-team_1_builder.add_edge("team_1_agent_2", "team_1_supervisor")
-
 team_1_graph = team_1_builder.compile()

 # define team 2 (same as the single supervisor example above)
@@ -179,31 +285,22 @@ team_2_graph = team_2_builder.compile()

 # define top-level supervisor

-class TopLevelState(MessagesState):
-    next: Literal["team_1", "team_2", "__end__"]
-
-builder = StateGraph(TopLevelState)
-def top_level_supervisor(state: TopLevelState):
+builder = StateGraph(MessagesState)
+def top_level_supervisor(state: MessagesState):
    # you can pass relevant parts of the state to the LLM (e.g., state["messages"])
    # to determine which team to call next. a common pattern is to call the model
    # with a structured output (e.g. force it to return an output with a "next_team" field)
    response = model.invoke(...)
-    # the "next" key will be used by the conditional edges to route execution
-    # to the appropriate team
-    return {"next": response["next_team"]}
+    # route to one of the teams or exit based on the supervisor's decision
+    # if the supervisor returns "__end__", the graph will finish execution
+    return Command(goto=response["next_team"])

-builder = StateGraph(TopLevelState)
+builder = StateGraph(MessagesState)
 builder.add_node(top_level_supervisor)
 builder.add_node(team_1_graph)
 builder.add_node(team_2_graph)

 builder.add_edge(START, "top_level_supervisor")
-# route to one of the teams or exit based on the supervisor's decision
-# if the top-level supervisor returns "__end__", the graph will finish execution
-builder.add_conditional_edges("top_level_supervisor", lambda state: state["next"])
-builder.add_edge("team_1_graph", "top_level_supervisor")
-builder.add_edge("team_2_graph", "top_level_supervisor")
-
 graph = builder.compile()
 ```

@@ -213,7 +310,7 @@ In this architecture we add individual agents as graph nodes and define the orde

 - **Explicit control flow (normal edges)**: LangGraph allows you to explicitly define the control flow of your application (i.e. the sequence of how agents communicate) explicitly, via [normal graph edges](./low_level.md#normal-edges). This is the most deterministic variant of this architecture above — we always know which agent will be called next ahead of time.

- **Dynamic control flow (conditional edges)**: in LangGraph you can allow LLMs to decide parts of your application control flow. This can be achieved by using [conditional edges](./low_level.md#conditional-edges). A special case of this is a [supervisor tool-calling](#supervisor-tool-calling) architecture. In that case, the tool-calling LLM powering the supervisor agent will make decisions about the order in which the tools (agents) are being called.
+- **Dynamic control flow (Command)**: in LangGraph you can allow LLMs to decide parts of your application control flow. This can be achieved by using [`Command`](./low_level.md#command). A special case of this is a [supervisor tool-calling](#supervisor-tool-calling) architecture. In that case, the tool-calling LLM powering the supervisor agent will make decisions about the order in which the tools (agents) are being called.

 ```python
 from langchain_openai import ChatOpenAI
@@ -218,13 +218,16 @@ The final thing you can optionally specify when calling `update_state` is `as_no

 ## Memory Store

-![Update](img/persistence/shared_state.png)
+![Model of shared state](img/persistence/shared_state.png)

 A [state schema](low_level.md#schema) specifies a set of keys that are populated as a graph is executed. As discussed above, state can be written by a checkpointer to a thread at each graph step, enabling state persistence.

 But, what if we want to retrain some information *across threads*? Consider the case of a chatbot where we want to retain specific information about the user across *all* chat conversations (e.g., threads) with that user!

-With checkpointers alone, we cannot share information across threads. This motivates the need for the `Store` interface. As an illustration, we can define an `InMemoryStore` to store information about a user across threads. We simply compile our graph with a checkpointer, as before, and will our new `in_memory_store`.
+With checkpointers alone, we cannot share information across threads. This motivates the need for the [`Store`](../reference/store.md#langgraph.store.base.BaseStore) interface. As an illustration, we can define an `InMemoryStore` to store information about a user across threads. We simply compile our graph with a checkpointer, as before, and with our new `in_memory_store` variable.
+
+### Basic Usage
+
 First, let's showcase this in isolation without using LangGraph.

 ```python
@@ -239,7 +242,7 @@ user_id = "1"
 namespace_for_memory = (user_id, "memories")
 ```

-We use the `store.put` to save memories to our namespace in the store. When we do this, we specify the namespace, as defined above, and a key-value pair for the memory: the key is simply a unique identifier for the memory (`memory_id`) and the value (a dictionary) is the memory itself.
+We use the `store.put` method to save memories to our namespace in the store. When we do this, we specify the namespace, as defined above, and a key-value pair for the memory: the key is simply a unique identifier for the memory (`memory_id`) and the value (a dictionary) is the memory itself.

 ```python
 memory_id = str(uuid.uuid4())
@@ -247,7 +250,7 @@ memory = {"food_preference" : "I like pizza"}
 in_memory_store.put(namespace_for_memory, memory_id, memory)
 ```

-We can read out memories in our namespace using `store.search`, which will return all memories for a given user as a list. The most recent memory is the last in the list.
+We can read out memories in our namespace using the `store.search` method, which will return all memories for a given user as a list. The most recent memory is the last in the list.

 ```python
 memories = in_memory_store.search(namespace_for_memory)
@@ -259,16 +262,69 @@ memories[-1].dict()
 'updated_at': '2024-10-02T17:22:31.590605+00:00'}
 ```

-Each memory type is a Python class with certain attributes. We can access it as a dictionary by converting via `.dict` as above.
+Each memory type is a Python class ([`Item`](https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.Item)) with certain attributes. We can access it as a dictionary by converting via `.dict` as above.
 The attributes it has are:

 - `value`: The value (itself a dictionary) of this memory
- `key`: The UUID for this memory in this namespace
+- `key`: A unique key for this memory in this namespace
 - `namespace`: A list of strings, the namespace of this memory type
 - `created_at`: Timestamp for when this memory was created
 - `updated_at`: Timestamp for when this memory was updated

-With this all in place, we use the `in_memory_store` in LangGraph. The `in_memory_store` works hand-in-hand with the checkpointer: the checkpointer saves state to threads, as discussed above, and the the `in_memory_store` allows us to store arbitrary information for access *across* threads. We compile the graph with both the checkpointer and the `in_memory_store` as follows. 
+### Semantic Search
+
+Beyond simple retrieval, the store also supports semantic search, allowing you to find memories based on meaning rather than exact matches. To enable this, configure the store with an embedding model:
+
+```python
+from langchain.embeddings import init_embeddings
+
+store = InMemoryStore(
+    index={
+        "embed": init_embeddings("openai:text-embedding-3-small"),  # Embedding provider
+        "dims": 1536,                              # Embedding dimensions
+        "fields": ["food_preference", "$"]              # Fields to embed
+    }
+)
+```
+
+Now when searching, you can use natural language queries to find relevant memories:
+
+```python
+# Find memories about food preferences
+# (This can be done after putting memories into the store)
+memories = store.search(
+    namespace_for_memory,
+    query="What does the user like to eat?",
+    limit=3  # Return top 3 matches
+)
+```
+
+You can control which parts of your memories get embedded by configuring the `fields` parameter or by specifying the `index` parameter when storing memories:
+
+```python
+# Store with specific fields to embed
+store.put(
+    namespace_for_memory,
+    str(uuid.uuid4()),
+    {
+        "food_preference": "I love Italian cuisine",
+        "context": "Discussing dinner plans"
+    },
+    index=["food_preference"]  # Only embed "food_preferences" field
+)
+
+# Store without embedding (still retrievable, but not searchable)
+store.put(
+    namespace_for_memory,
+    str(uuid.uuid4()),
+    {"system_info": "Last updated: 2024-01-01"},
+    index=False
+)
+```
+
+### Using in LangGraph
+
+With this all in place, we use the `in_memory_store` in LangGraph. The `in_memory_store` works hand-in-hand with the checkpointer: the checkpointer saves state to threads, as discussed above, and the `in_memory_store` allows us to store arbitrary information for access *across* threads. We compile the graph with both the checkpointer and the `in_memory_store` as follows. 

 ```python
 from langgraph.checkpoint.memory import MemorySaver
@@ -296,7 +352,7 @@ for update in graph.stream(
    print(update)
 ```

-We can access the `in_memory_store` and the `user_id` in *any node* by passing `store: BaseStore` and `config: RunnableConfig` as node arguments. Just as we saw above, simply use the `put` method to save memories to the store.
+We can access the `in_memory_store` and the `user_id` in *any node* by passing `store: BaseStore` and `config: RunnableConfig` as node arguments. Here's how we might use semantic search in a node to find relevant memories:

 ```python
 def update_memory(state: MessagesState, config: RunnableConfig, *, store: BaseStore):
@@ -317,7 +373,7 @@ def update_memory(state: MessagesState, config: RunnableConfig, *, store: BaseSt

 ```

-As we showed above, we can also access the store in any node and use `search` to get memories. Recall the the memories are returned as a list of objects that can be converted to a dictionary.
+As we showed above, we can also access the store in any node and use the `store.search` method to get memories. Recall the the memories are returned as a list of objects that can be converted to a dictionary.

 ```python
 memories[-1].dict()
@@ -332,12 +388,15 @@ We can access the memories and use them in our model call.

 ```python
 def call_model(state: MessagesState, config: RunnableConfig, *, store: BaseStore):
-    
    # Get the user id from the config
    user_id = config["configurable"]["user_id"]
    
-    # Get the memories for the user from the store
-    memories = store.search(("memories", user_id))
+    # Search based on the most recent message
+    memories = store.search(
+        namespace,
+        query=state["messages"][-1].content,
+        limit=3
+    )
    info = "\n".join([d.value["memory"] for d in memories])
    
    # ... Use memories in the model call
@@ -356,7 +415,22 @@ for update in graph.stream(
    print(update)
 ```

-When we use the LangGraph API, either locally (e.g., in LangGraph Studio) or with LangGraph Cloud, the memory store is available to use by default and does not need to be specified during graph compilation.
+When we use the LangGraph Platform, either locally (e.g., in LangGraph Studio) or with LangGraph Cloud, the base store is available to use by default and does not need to be specified during graph compilation. To enable semantic search, however, you **do** need to configure the indexing settings in your `langgraph.json` file. For example:
+
+```json
+{
+    ...
+    "store": {
+        "index": {
+            "embed": "openai:text-embeddings-3-small",
+            "dims": 1536,
+            "fields": ["$"]
+        }
+    }
+}
+```
+
+See the [deployment guide](../cloud/deployment/semantic_search.md) for more details and configuration options.

 ## Checkpointer libraries

@@ -405,4 +479,4 @@ Lastly, checkpointing also provides fault-tolerance and error recovery: if one o

 #### Pending writes

-Additionally, when a graph node fails mid-execution at a given superstep, LangGraph stores pending checkpoint writes from any other nodes that completed successfully at that superstep, so that whenever we resume graph execution from that superstep we don't re-run the successful nodes.
+Additionally, when a graph node fails mid-execution at a given superstep, LangGraph stores pending checkpoint writes from any other nodes that completed successfully at that superstep, so that whenever we resume graph execution from that superstep we don't re-run the successful nodes.
@@ -7,7 +7,7 @@

 ## Versions

-There are two versions of the self hosted deployment: [Self-Hosted Enterprise](./deployment_options.md#self-hosted-enterprise) and [Self-Hosted Lite](./deployment_options.md#self-hosted-lite).
+There are two versions of the self-hosted deployment: [Self-Hosted Enterprise](./deployment_options.md#self-hosted-enterprise) and [Self-Hosted Lite](./deployment_options.md#self-hosted-lite).

 ### Self-Hosted Lite

@@ -34,6 +34,10 @@ To use the Self-Hosted Enterprise version, you must acquire a license key that y

 For step-by-step instructions, see [How to set up a self-hosted deployment of LangGraph](../how-tos/deploy-self-hosted.md).

+## Helm Chart
+
+If you would like to deploy LangGraph Cloud on Kubernetes, you can use this [Helm chart](https://github.com/langchain-ai/helm/blob/main/charts/langgraph-cloud/README.md).
+
 ## Related

 - [How to set up a self-hosted deployment of LangGraph](../how-tos/deploy-self-hosted.md).
@@ -1,27 +1,62 @@
 # Template Applications

-!!! note Prerequisites
-
-    - [LangGraph Studio](./langgraph_studio.md)
-
 Templates are open source reference applications designed to help you get started quickly when building with LangGraph. They provide working examples of common agentic workflows that can be customized to your needs.

-Templates can be accessed via [LangGraph Studio](langgraph_studio.md), or cloned directly from Github. You can download LangGraph Studio and see available templates [here](https://studio.langchain.com/).
+You can create an application from a template using the LangGraph CLI.

-## Available templates
+!!! info "Requirements"

- **New LangGraph Project**: A simple, minimal chatbot with memory.
-  - [Python](https://github.com/langchain-ai/new-langgraph-project)
-  - [JS/TS](https://github.com/langchain-ai/new-langgraphjs-project)
- **ReAct Agent**: A simple agent that can be flexibly extended to many tools.
-  - [Python](https://github.com/langchain-ai/react-agent)
-  - [JS/TS](https://github.com/langchain-ai/react-agent-js)
- **Memory Agent**: A ReAct-style agent with an additional tool to store memories for use across conversational threads.
-  - [Python](https://github.com/langchain-ai/memory-agent)
-  - [JS/TS](https://github.com/langchain-ai/memory-agent-js)
- **Retrieval Agent**: An agent that includes a retrieval-based question-answering system.
-  - [Python](https://github.com/langchain-ai/retrieval-agent-template)
-  - [JS/TS](https://github.com/langchain-ai/retrieval-agent-template-js)
- **Data-enrichment Agent**: An agent that performs web searches and organizes its findings into a structured format.
-  - [Python](https://github.com/langchain-ai/data-enrichment)
-  - [JS/TS](https://github.com/langchain-ai/data-enrichment-js)
+    - Python >= 3.11
+    - [LangGraph CLI](https://langchain-ai.github.io/langgraph/cloud/reference/cli/): Requires langchain-cli[inmem] >= 0.1.58
+
+## Install the LangGraph CLI
+
+```bash
+pip install "langgraph-cli[inmem]==0.1.58" python-dotenv
+```
+
+## Available Templates
+
+| Template                  | Description                                                                              | Python                                                           | JS/TS                                                               |
+|---------------------------|------------------------------------------------------------------------------------------|------------------------------------------------------------------|---------------------------------------------------------------------|
+| **New LangGraph Project** | A simple, minimal chatbot with memory.                                                   | [Repo](https://github.com/langchain-ai/new-langgraph-project)    | [Repo](https://github.com/langchain-ai/new-langgraphjs-project)     |
+| **ReAct Agent**           | A simple agent that can be flexibly extended to many tools.                              | [Repo](https://github.com/langchain-ai/react-agent)              | [Repo](https://github.com/langchain-ai/react-agent-js)              |
+| **Memory Agent**          | A ReAct-style agent with an additional tool to store memories for use across threads.    | [Repo](https://github.com/langchain-ai/memory-agent)             | [Repo](https://github.com/langchain-ai/memory-agent-js)             |
+| **Retrieval Agent**       | An agent that includes a retrieval-based question-answering system.                      | [Repo](https://github.com/langchain-ai/retrieval-agent-template) | [Repo](https://github.com/langchain-ai/retrieval-agent-template-js) |
+| **Data-Enrichment Agent** | An agent that performs web searches and organizes its findings into a structured format. | [Repo](https://github.com/langchain-ai/data-enrichment)          | [Repo](https://github.com/langchain-ai/data-enrichment-js)          |
+
+
+## 🌱 Create a LangGraph App
+
+To create a new app from a template, use the `langgraph new` command.
+
+```bash
+langgraph new
+```
+
+## Next Steps
+
+Review the `README.md` file in the root of your new LangGraph app for more information about the template and how to customize it.
+
+After configuring the app properly and adding your API keys, you can start the app using the LangGraph CLI:
+
+```bash
+langgraph dev 
+```
+
+See the following guides for more information on how to deploy your app:
+
+- **[Launch Local LangGraph Server](../tutorials/langgraph-platform/local-server.md)**: This quick start guide shows how to start a LangGraph Server locally for the **ReAct Agent** template. The steps are similar for other templates.
+- **[Deploy to LangGraph Cloud](../cloud/quick_start.md)**: Deploy your LangGraph app using LangGraph Cloud.
+ 
+### LangGraph Framework
+
+- **[LangGraph Concepts](../concepts/index.md)**: Learn the foundational concepts of LangGraph.
+- **[LangGraph How-to Guides](../how-tos/index.md)**: Guides for common tasks with LangGraph.
+
+### 📚 Learn More about LangGraph Platform
+
+Expand your knowledge with these resources:
+
+- **[LangGraph Platform Concepts](../concepts/index.md#langgraph-platform)**: Understand the foundational concepts of the LangGraph Platform.
+- **[LangGraph Platform How-to Guides](../how-tos/index.md#langgraph-platform)**: Discover step-by-step guides to build and deploy applications.
@@ -0,0 +1,171 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "8381b6e0-29a6-48c5-b451-5d2549351249",
+   "metadata": {},
+   "source": [
+    "# How to use LangGraph Platform to deploy CrewAI, AutoGen, and other frameworks\n",
+    "\n",
+    "[LangGraph Platform](https://langchain-ai.github.io/langgraph/concepts/langgraph_platform/) provides infrastructure for deploying agents. This integrates seamlessly with LangGraph, but can also work with other frameworks. The way to make this work is to wrap the agent in a single LangGraph node, and have that be the entire graph.\n",
+    "\n",
+    "Doing so will allow you to deploy to LangGraph Platform, and allows you to get a lot of the [benefits](https://langchain-ai.github.io/langgraph/concepts/langgraph_platform/). You get horizontally scalable infrastructure, a task queue to handle bursty operations, a persistence layer to power short term memory, and long term memory support.\n",
+    "\n",
+    "In this guide we show how to do this with an AutoGen agent, but this method should work for agents defined in other frameworks like CrewAI, LlamaIndex, and others as well."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1113cb16-b538-448c-924c-85731ce96ebd",
+   "metadata": {},
+   "source": [
+    "## Setup"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "f05993fa-9d03-4f45-bc13-0a8d87260d86",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [],
+   "source": [
+    "%pip install autogen langgraph"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f4e0ca12-1714-4776-a30a-9527e519799b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import getpass\n",
+    "import os\n",
+    "\n",
+    "\n",
+    "def _set_env(var: str):\n",
+    "    if not os.environ.get(var):\n",
+    "        os.environ[var] = getpass.getpass(f\"{var}: \")\n",
+    "\n",
+    "\n",
+    "_set_env(\"OPENAI_API_KEY\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1926bbc3-6b06-41e0-9604-860a2bbf8fa3",
+   "metadata": {},
+   "source": [
+    "## Define autogen agent\n",
+    "\n",
+    "Here we define our AutoGen agent. From https://github.com/microsoft/autogen/blob/0.2/notebook/agentchat_web_info.ipynb"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d4a14dc7-d565-4207-8788-525f85b9fb27",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import autogen\n",
+    "import os\n",
+    "\n",
+    "config_list = [{\"model\": \"gpt-4o\", \"api_key\": os.environ[\"OPENAI_API_KEY\"]}]\n",
+    "\n",
+    "llm_config = {\n",
+    "    \"timeout\": 600,\n",
+    "    \"cache_seed\": 42,\n",
+    "    \"config_list\": config_list,\n",
+    "    \"temperature\": 0,\n",
+    "}\n",
+    "\n",
+    "autogen_agent = autogen.AssistantAgent(\n",
+    "    name=\"assistant\",\n",
+    "    llm_config=llm_config,\n",
+    ")\n",
+    "\n",
+    "user_proxy = autogen.UserProxyAgent(\n",
+    "    name=\"user_proxy\",\n",
+    "    human_input_mode=\"NEVER\",\n",
+    "    max_consecutive_auto_reply=10,\n",
+    "    is_termination_msg=lambda x: x.get(\"content\", \"\").rstrip().endswith(\"TERMINATE\"),\n",
+    "    code_execution_config={\n",
+    "        \"work_dir\": \"web\",\n",
+    "        \"use_docker\": False,\n",
+    "    },  # Please set use_docker=True if docker is available to run the generated code. Using docker is safer than running the generated code directly.\n",
+    "    llm_config=llm_config,\n",
+    "    system_message=\"Reply TERMINATE if the task has been solved at full satisfaction. Otherwise, reply CONTINUE, or the reason why the task is not solved yet.\",\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b1170836-f23e-4e4c-ab83-ce791cd7fbd2",
+   "metadata": {},
+   "source": [
+    "## Wrap in LangGraph\n",
+    "\n",
+    "We now wrap the AutoGen agent in a single LangGraph node, and make that the entire graph.\n",
+    "The main thing this involves is defining an Input and Output schema for the node, which you would need to do if deploying this manually, so it's no extra work"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "7b417c16-ff4e-4d5c-a9a9-0aaeeef6ede5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langgraph.graph import StateGraph, MessagesState\n",
+    "\n",
+    "\n",
+    "def call_autogen_agent(state: MessagesState):\n",
+    "    last_message = state[\"messages\"][-1]\n",
+    "    response = user_proxy.initiate_chat(autogen_agent, message=last_message.content)\n",
+    "    # get the final response from the agent\n",
+    "    content = response.chat_history[-1][\"content\"]\n",
+    "    return {\"messages\": {\"role\": \"assistant\", \"content\": content}}\n",
+    "\n",
+    "\n",
+    "graph = StateGraph(MessagesState)\n",
+    "graph.add_node(call_autogen_agent)\n",
+    "graph.set_entry_point(\"call_autogen_agent\")\n",
+    "graph = graph.compile()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f6a18377-ac29-478f-a76a-b213f1a3c85d",
+   "metadata": {},
+   "source": [
+    "## Deploy with LangGraph Platform\n",
+    "\n",
+    "You can now deploy this as you normally would with LangGraph Platform. See [these instructions](https://langchain-ai.github.io/langgraph/concepts/deployment_options/) for more details."
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
@@ -345,7 +345,7 @@
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
-   "version": "3.11.4"
+   "version": "3.11.1"
  }
 },
 "nbformat": 4,
@@ -41,6 +41,9 @@
    "    <p>\n",
    "    Support for the <code><a href=\"https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.BaseStore\">Store</a></code> API that is used in this guide was added in LangGraph <code>v0.2.32</code>.\n",
    "    </p>\n",
+    "    <p>\n",
+    "    Support for <b>index</b> and <b>query</b> arguments of the <code><a href=\"https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.BaseStore\">Store</a></code> API that is used in this guide was added in LangGraph <code>v0.2.54</code>.\n",
+    "    </p>\n",
    "</div>\n",
    "\n",
    "## Setup\n",
@@ -114,7 +117,7 @@
    "\n",
    "Importantly, to determine the user, we will be passing `user_id` via the config keyword argument of the node function.\n",
    "\n",
-    "Let's first define an `InMemoryStore` which is already populated with some memories about the users."
+    "Let's first define an `InMemoryStore` already populated with some memories about the users."
   ]
  },
  {
@@ -125,8 +128,14 @@
   "outputs": [],
   "source": [
    "from langgraph.store.memory import InMemoryStore\n",
+    "from langchain_openai import OpenAIEmbeddings\n",
    "\n",
-    "in_memory_store = InMemoryStore()"
+    "in_memory_store = InMemoryStore(\n",
+    "    index={\n",
+    "        \"embed\": OpenAIEmbeddings(model=\"text-embedding-3-small\"),\n",
+    "        \"dims\": 1536,\n",
+    "    }\n",
+    ")"
   ]
  },
  {
@@ -163,7 +172,7 @@
    "def call_model(state: MessagesState, config: RunnableConfig, *, store: BaseStore):\n",
    "    user_id = config[\"configurable\"][\"user_id\"]\n",
    "    namespace = (\"memories\", user_id)\n",
-    "    memories = store.search(namespace)\n",
+    "    memories = store.search(namespace, query=str(state[\"messages\"][-1].content))\n",
    "    info = \"\\n\".join([d.value[\"data\"] for d in memories])\n",
    "    system_msg = f\"You are a helpful assistant talking to the user. User info: {info}\"\n",
    "\n",
@@ -17,14 +17,18 @@ You will need to do the following:
 2. Build a docker image with the [LangGraph Server](../concepts/langgraph_server.md) using the [LangGraph CLI](../concepts/langgraph_cli.md).
 3. Deploy a web server that will run the docker image and pass in the necessary environment variables.

+## Helm Chart
+
+If you would like to deploy LangGraph Cloud on Kubernetes, you can use this [Helm chart](https://github.com/langchain-ai/helm/blob/main/charts/langgraph-cloud/README.md).
+
 ## Environment Variables

 You will eventually need to pass in the following environment variables to the LangGraph Deploy server:

 - `REDIS_URI`: Connection details to a Redis instance. Redis will be used as a pub-sub broker to enable streaming real time output from background runs.
 - `DATABASE_URI`: Postgres connection details. Postgres will be used to store assistants, threads, runs, persist thread state and long term memory, and to manage the state of the background task queue with 'exactly once' semantics.
- `LANGSMITH_API_KEY`: (If using [Self-Hosted Lite]) LangSmith API key. This will be used to authenticate ONCE at server start up.
- `LANGGRAPH_CLOUD_LICENSE_KEY`: (If using Self-Hosted Enterprise) LangGraph Platform license key. This will be used to authenticate ONCE at server start up.
+- `LANGSMITH_API_KEY`: (If using [Self-Hosted Lite](../concepts/deployment_options.md#self-hosted-lite)) LangSmith API key. This will be used to authenticate ONCE at server start up.
+- `LANGGRAPH_CLOUD_LICENSE_KEY`: (If using [Self-Hosted Enterprise](../concepts/deployment_options.md#self-hosted-enterprise)) LangGraph Platform license key. This will be used to authenticate ONCE at server start up.


 ## Build the Docker Image
@@ -70,7 +74,7 @@ If you want to run this quickly without setting up a separate Redis and Postgres
    * You need to replace `my-image` with the name of the image you built in the previous step (from `langgraph build`).
    and you should provide appropriate values for `REDIS_URI`, `DATABASE_URI`, and `LANGSMITH_API_KEY`.
    * If your application requires additional environment variables, you can pass them in a similar way.
-    * If using Self-Hosted Enterprise, you must provide `LANGGRAPH_CLOUD_LICENSE_KEY` as an additional environment variable.
+    * If using [Self-Hosted Enterprise](../concepts/deployment_options.md#self-hosted-enterprise), you must provide `LANGGRAPH_CLOUD_LICENSE_KEY` as an additional environment variable.


 ### Using Docker Compose
@@ -20,6 +20,7 @@ These how-to guides show how to achieve that controllability.
 - [How to create branches for parallel execution](branching.ipynb)
 - [How to create map-reduce branches for parallel execution](map-reduce.ipynb)
 - [How to control graph recursion limit](recursion-limit.ipynb)
+- [How to combine control flow and state updates with Command](command.ipynb)

 ### Persistence

@@ -39,6 +40,8 @@ LangGraph makes it easy to manage conversation [memory](../concepts/memory.md) i
 - [How to manage conversation history](memory/manage-conversation-history.ipynb)
 - [How to delete messages](memory/delete-messages.ipynb)
 - [How to add summary conversation memory](memory/add-summary-conversation-history.ipynb)
+- [How to add long-term memory (cross-thread)](cross-thread-persistence.ipynb)
+- [How to use semantic search for long-term memory](memory/semantic-search.ipynb)

 ### Human-in-the-loop

@@ -70,7 +73,7 @@ you to involve humans in the decision-making process of your graph. These how-to

 ### Tool calling

-[Tool calling](https://python.langchain.com/docs/concepts/tool_calling/) is a type of chat model API that accepts tool schemas, along with messages, as input and returns invocations of those tools as part of the output message. 
+[Tool calling](https://python.langchain.com/docs/concepts/tool_calling/) is a type of chat model API that accepts tool schemas, along with messages, as input and returns invocations of those tools as part of the output message.

 These how-to guides show common patterns for tool calling with LangGraph:

@@ -78,6 +81,7 @@ These how-to guides show common patterns for tool calling with LangGraph:
 - [How to handle tool calling errors](tool-calling-errors.ipynb)
 - [How to pass runtime values to tools](pass-run-time-values-to-tools.ipynb)
 - [How to pass config to tools](pass-config-to-tools.ipynb)
+- [How to update graph state from tools](update-state-from-tools.ipynb)
 - [How to handle large numbers of tools](many-tools.ipynb)

 ### Subgraphs
@@ -88,6 +92,12 @@ These how-to guides show common patterns for tool calling with LangGraph:
 - [How to view and update state in subgraphs](subgraphs-manage-state.ipynb)
 - [How to transform inputs and outputs of a subgraph](subgraph-transform-state.ipynb)

+### Multi-agent
+
+- [How to build a multi-agent network](multi-agent-network.ipynb)
+
+See the [multi-agent tutorials](../tutorials/index.md#multi-agent-systems) for implementations of other multi-agent architectures.
+
 ### State Management

 - [How to use Pydantic model as state](state-model.ipynb)
@@ -103,6 +113,7 @@ These how-to guides show common patterns for tool calling with LangGraph:
 - [How to force function calling agent to structure output](react-agent-structured-output.ipynb)
 - [How to pass custom LangSmith run ID for graph runs](run-id-langsmith.ipynb)
 - [How to return state before hitting recursion limit](return-when-recursion-limit-hits.ipynb)
+- [How to integrate LangGraph with AutoGen, CrewAI, and other frameworks](autogen-integration.ipynb)

 ### Prebuilt ReAct Agent

@@ -117,12 +128,13 @@ These guides show how to use the prebuilt ReAct agent:
 - [How to add a custom system prompt to a ReAct agent](create-react-agent-system-prompt.ipynb)
 - [How to add human-in-the-loop processes to a ReAct agent](create-react-agent-hitl.ipynb)
 - [How to create prebuilt ReAct agent from scratch](react-agent-from-scratch.ipynb)
+- [How to add semantic search for long-term memory to a ReAct agent](memory/semantic-search.ipynb#using-in-create-react-agent)

 ## LangGraph Platform

 This section includes how-to guides for LangGraph Platform.

-LangGraph Platform is a commercial solution for deploying agentic applications in production, built on the open-source LangGraph framework. 
+LangGraph Platform is a commercial solution for deploying agentic applications in production, built on the open-source LangGraph framework.

 The LangGraph Platform offers a few different deployment options described in the [deployment options guide](../concepts/deployment_options.md).

@@ -138,9 +150,11 @@ Learn how to set up your app for deployment to LangGraph Platform:
 - [How to set up app for deployment (requirements.txt)](../cloud/deployment/setup.md)
 - [How to set up app for deployment (pyproject.toml)](../cloud/deployment/setup_pyproject.md)
 - [How to set up app for deployment (JavaScript)](../cloud/deployment/setup_javascript.md)
+- [How to add semantic search](../cloud/deployment/semantic_search.md)
 - [How to customize Dockerfile](../cloud/deployment/custom_docker.md)
 - [How to test locally](../cloud/deployment/test_locally.md)
 - [How to rebuild graph at runtime](../cloud/deployment/graph_rebuild.md)
+- [How to use LangGraph Platform to deploy CrewAI, AutoGen, and other frameworks](autogen-langgraph-platform.ipynb)

 ### Deployment

@@ -148,7 +162,8 @@ LangGraph applications can be deployed using LangGraph Cloud, which provides a r

 - [How to deploy to LangGraph cloud](../cloud/deployment/cloud.md)
 - [How to deploy to a self-hosted environment](./deploy-self-hosted.md)
- [How to interact with the deployment using RemoteGraph](./use-remote-graph.md) 
+- [How to interact with the deployment using RemoteGraph](./use-remote-graph.md)
+
 ### Assistants

 [Assistants](../concepts/assistants.md) is a configured instance of a template.
@@ -163,7 +178,7 @@ LangGraph applications can be deployed using LangGraph Cloud, which provides a r

 ### Runs

-LangGraph Cloud supports multiple types of runs besides streaming runs.
+LangGraph Platform supports multiple types of runs besides streaming runs.

 - [How to run an agent in the background](../cloud/how-tos/background_run.md)
 - [How to run multiple agents in the same thread](../cloud/how-tos/same-thread.md)
@@ -183,7 +198,7 @@ Streaming the results of your LLM application is vital for ensuring a good user

 ### Human-in-the-loop

-When creating complex graphs, leaving every decision up to the LLM can be dangerous, especially when the decisions involve invoking certain tools or accessing specific documents. To remedy this, LangGraph allows you to insert human-in-the-loop behavior to ensure your graph does not have undesired outcomes. Read more about the different ways you can add human-in-the-loop capabilities to your LangGraph Cloud projects in these how-to guides:
+When designing complex graphs, relying entirely on the LLM for decision-making can be risky, particularly when it involves tools that interact with files, APIs, or databases. These interactions may lead to unintended data access or modifications, depending on the use case. To mitigate these risks, LangGraph allows you to integrate human-in-the-loop behavior, ensuring your LLM applications operate as intended without undesirable outcomes.

 - [How to add a breakpoint](../cloud/how-tos/human_in_the_loop_breakpoint.md)
 - [How to wait for user input](../cloud/how-tos/human_in_the_loop_user_input.md)
@@ -193,7 +208,7 @@ When creating complex graphs, leaving every decision up to the LLM can be danger

 ### Double-texting

-Graph execution can take a while, and sometimes users may change their mind about the input they wanted to send before their original input has finished running. For example, a user might notice a typo in their original request and will edit the prompt and resend it. Deciding what to do in these cases is important for ensuring a smooth user experience and preventing your graphs from behaving in unexpected ways. The following how-to guides provide information on the various options LangGraph Cloud gives you for dealing with double-texting:
+Graph execution can take a while, and sometimes users may change their mind about the input they wanted to send before their original input has finished running. For example, a user might notice a typo in their original request and will edit the prompt and resend it. Deciding what to do in these cases is important for ensuring a smooth user experience and preventing your graphs from behaving in unexpected ways.

 - [How to use the interrupt option](../cloud/how-tos/interrupt_concurrent.md)
 - [How to use the rollback option](../cloud/how-tos/rollback_concurrent.md)
@@ -213,8 +228,9 @@ Graph execution can take a while, and sometimes users may change their mind abou
 LangGraph Studio is a built-in UI for visualizing, testing, and debugging your agents.

 - [How to connect to a LangGraph Cloud deployment](../cloud/how-tos/test_deployment.md)
- [How to connect to a local deployment](../cloud/how-tos/test_local_deployment.md)
- [How to test your graph in LangGraph Studio](../cloud/how-tos/invoke_studio.md)
+- [How to connect to a local dev server](../how-tos/local-studio.md)
+- [How to connect to a local deployment (Docker)](../cloud/how-tos/test_local_deployment.md)
+- [How to test your graph in LangGraph Studio (MacOS only)](../cloud/how-tos/invoke_studio.md)
 - [How to interact with threads in LangGraph Studio](../cloud/how-tos/threads_studio.md)

 ## Troubleshooting
@@ -226,5 +242,3 @@ These are the guides for resolving common errors you may find while building wit
 - [INVALID_GRAPH_NODE_RETURN_VALUE](../troubleshooting/errors/INVALID_GRAPH_NODE_RETURN_VALUE.md)
 - [MULTIPLE_SUBGRAPHS](../troubleshooting/errors/MULTIPLE_SUBGRAPHS.md)
 - [INVALID_CHAT_HISTORY](../troubleshooting/errors/INVALID_CHAT_HISTORY.md)
-
-
@@ -0,0 +1,86 @@
+# How to connect a local agent to LangGraph Studio
+
+This guide shows you how to connect your local agent to [LangGraph Studio](../concepts/langgraph_studio.md) for visualization, interaction, and debugging.
+
+## Connection Options
+
+There are two ways to connect your local agent to LangGraph Studio:
+
+- [LangGraph Desktop](../concepts/langgraph_studio.md#desktop-app): Application, Mac only, requires Docker
+- [Development Server](../concepts/langgraph_studio.md#dev-server): Python package, all platforms, no Docker
+
+In this guide we will cover how to use the development server as that is generally an easier and better experience.
+
+## Setup your application
+
+First, you will need to setup your application in the proper format.
+This means defining a `langgraph.json` file which contains paths to your agent(s).
+See [this guide](../concepts/application_structure.md) for information on how to do so.
+
+## Install langgraph-cli
+
+You will need to install [`langgraph-cli`](../cloud/reference/cli.md#langgraph-cli) (version `0.1.55` or higher).
+You will need to make sure to install the `inmem` extras.
+
+```shell
+pip install "langgraph-cli[inmem]==0.1.55"
+```
+
+## Run the development server
+
+1. Navigate to your project directory (where `langgraph.json` is located)
+
+2. Start the server:
+   ```bash
+   langgraph dev
+   ```
+
+This will look for the `langgraph.json` file in your current directory. 
+In there, it will find the paths to the graph(s), and start those up.
+It will then automatically connect to the cloud-hosted studio.
+
+## Use the studio
+
+After connecting to the studio, a browser window should automatically pop up.
+This will use the cloud hosted studio UI to connect to your local development server.
+Your graph is still running locally, the UI is connecting to visualizing the agent and threads that are defined locally.
+
+The graph will always use the most up-to-date code, so you will be able to change the underlying code and have it automatically reflected in the studio.
+This is useful for debugging workflows.
+You can run your graph in the UI until it messes up, go in and change your code, and then rerun from the node that failed.
+
+# (Optional) Attach a debugger
+
+For step-by-step debugging with breakpoints and variable inspection:
+
+```bash
+# Install debugpy package
+pip install debugpy
+
+# Start server with debugging enabled
+langgraph dev --debug-port 5678
+```
+
+Then attach your preferred debugger:
+
+=== "VS Code"
+    Add this configuration to `launch.json`:
+    ```json
+    {
+      "name": "Attach to LangGraph",
+      "type": "debugpy",
+      "request": "attach",
+      "connect": {
+        "host": "0.0.0.0",
+        "port": 5678
+      }
+    }
+    ```
+    Specify the port number you chose in the previous step.
+
+=== "PyCharm"
+    1. Go to Run → Edit Configurations
+    2. Click + and select "Python Debug Server"
+    3. Set IDE host name: `localhost`
+    4. Set port: `5678` (or the port number you chose in the previous step)
+    5. Click "OK" and start debugging
@@ -0,0 +1,532 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# How to add semantic search to your agent's memory\n",
+    "\n",
+    "This guide shows how to enable semantic search in your agent's memory store. This lets search for items in the store by semantic similarity.\n",
+    "\n",
+    "!!! tip Prerequisites\n",
+    "    This guide assumes familiarity with the [memory in LangGraph](https://langchain-ai.github.io/langgraph/concepts/memory/).\n",
+    "\n",
+    "First, install this guide's prerequisites."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%capture --no-stderr\n",
+    "%pip install -U langgraph langchain-openai langchain"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import getpass\n",
+    "import os\n",
+    "\n",
+    "\n",
+    "def _set_env(var: str):\n",
+    "    if not os.environ.get(var):\n",
+    "        os.environ[var] = getpass.getpass(f\"{var}: \")\n",
+    "\n",
+    "\n",
+    "_set_env(\"OPENAI_API_KEY\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Next, create the store with an [index configuration](https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.IndexConfig). By default, stores are configured without semantic/vector search. You can opt in to indexing items when creating the store by providing an [IndexConfig](https://langchain-ai.github.io/langgraph/reference/store/#langgraph.store.base.IndexConfig) to the store's constructor. If your store class does not implement this interface, or if you do not pass in an index configuration, semantic search is disabled, and all `index` arguments passed to `put` or `aput` will have no effect. Below is an example."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/gf/6rnp_mbx5914kx7qmmh7xzmw0000gn/T/ipykernel_83572/2318027494.py:5: LangChainBetaWarning: The function `init_embeddings` is in beta. It is actively being worked on, so the API may change.\n",
+      "  embeddings = init_embeddings(\"openai:text-embedding-3-small\")\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain.embeddings import init_embeddings\n",
+    "from langgraph.store.memory import InMemoryStore\n",
+    "\n",
+    "# Create store with semantic search enabled\n",
+    "embeddings = init_embeddings(\"openai:text-embedding-3-small\")\n",
+    "store = InMemoryStore(\n",
+    "    index={\n",
+    "        \"embed\": embeddings,\n",
+    "        \"dims\": 1536,\n",
+    "    }\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now let's store some memories:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Store some memories\n",
+    "store.put((\"user_123\", \"memories\"), \"1\", {\"text\": \"I love pizza\"})\n",
+    "store.put((\"user_123\", \"memories\"), \"2\", {\"text\": \"I prefer Italian food\"})\n",
+    "store.put((\"user_123\", \"memories\"), \"3\", {\"text\": \"I don't like spicy food\"})\n",
+    "store.put((\"user_123\", \"memories\"), \"3\", {\"text\": \"I am studying econometrics\"})\n",
+    "store.put((\"user_123\", \"memories\"), \"3\", {\"text\": \"I am a plumber\"})"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Search memories using natural language:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Memory: I prefer Italian food (similarity: 0.46482669521168163)\n",
+      "Memory: I love pizza (similarity: 0.35514845174380766)\n",
+      "Memory: I am a plumber (similarity: 0.155698702336571)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Find memories about food preferences\n",
+    "memories = store.search((\"user_123\", \"memories\"), query=\"I like food?\", limit=5)\n",
+    "\n",
+    "for memory in memories:\n",
+    "    print(f'Memory: {memory.value[\"text\"]} (similarity: {memory.score})')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Using in your agent\n",
+    "\n",
+    "Add semantic search to any node by injecting the store."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "What are you in the mood for? Since you love Italian food and pizza, would you like to order a pizza or try making one at home?"
+     ]
+    }
+   ],
+   "source": [
+    "from typing import Optional\n",
+    "\n",
+    "from langchain.chat_models import init_chat_model\n",
+    "from langgraph.store.base import BaseStore\n",
+    "\n",
+    "from langgraph.graph import START, MessagesState, StateGraph\n",
+    "\n",
+    "llm = init_chat_model(\"openai:gpt-4o-mini\")\n",
+    "\n",
+    "\n",
+    "def chat(state, *, store: BaseStore):\n",
+    "    # Search based on user's last message\n",
+    "    items = store.search(\n",
+    "        (\"user_123\", \"memories\"), query=state[\"messages\"][-1].content, limit=2\n",
+    "    )\n",
+    "    memories = \"\\n\".join(item.value[\"text\"] for item in items)\n",
+    "    memories = f\"## Memories of user\\n{memories}\" if memories else \"\"\n",
+    "    response = llm.invoke(\n",
+    "        [\n",
+    "            {\"role\": \"system\", \"content\": f\"You are a helpful assistant.\\n{memories}\"},\n",
+    "            *state[\"messages\"],\n",
+    "        ]\n",
+    "    )\n",
+    "    return {\"messages\": [response]}\n",
+    "\n",
+    "\n",
+    "builder = StateGraph(MessagesState)\n",
+    "builder.add_node(chat)\n",
+    "builder.add_edge(START, \"chat\")\n",
+    "graph = builder.compile(store=store)\n",
+    "\n",
+    "for message, metadata in graph.stream(\n",
+    "    input={\"messages\": [{\"role\": \"user\", \"content\": \"I'm hungry\"}]},\n",
+    "    stream_mode=\"messages\",\n",
+    "):\n",
+    "    print(message.content, end=\"\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Using in `create_react_agent`\n",
+    "\n",
+    "Add semantic search to your tool calling agent by injecting the store in the `state_modifier`. You can also use the store in a tool to let your agent manually store or search for memories."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import uuid\n",
+    "from typing import Optional\n",
+    "\n",
+    "from langchain.chat_models import init_chat_model\n",
+    "from langchain_core.tools import InjectedToolArg\n",
+    "from langgraph.store.base import BaseStore\n",
+    "from typing_extensions import Annotated\n",
+    "\n",
+    "from langgraph.prebuilt import create_react_agent\n",
+    "\n",
+    "\n",
+    "def prepare_messages(state, *, store: BaseStore):\n",
+    "    # Search based on user's last message\n",
+    "    items = store.search(\n",
+    "        (\"user_123\", \"memories\"), query=state[\"messages\"][-1].content, limit=2\n",
+    "    )\n",
+    "    memories = \"\\n\".join(item.value[\"text\"] for item in items)\n",
+    "    memories = f\"## Memories of user\\n{memories}\" if memories else \"\"\n",
+    "    return [\n",
+    "        {\"role\": \"system\", \"content\": f\"You are a helpful assistant.\\n{memories}\"}\n",
+    "    ] + state[\"messages\"]\n",
+    "\n",
+    "\n",
+    "# You can also use the store directly within a tool!\n",
+    "def upsert_memory(\n",
+    "    content: str,\n",
+    "    *,\n",
+    "    memory_id: Optional[uuid.UUID] = None,\n",
+    "    store: Annotated[BaseStore, InjectedToolArg],\n",
+    "):\n",
+    "    \"\"\"Upsert a memory in the database.\"\"\"\n",
+    "    # The LLM can use this tool to store a new memory\n",
+    "    mem_id = memory_id or uuid.uuid4()\n",
+    "    store.put(\n",
+    "        (\"user_123\", \"memories\"),\n",
+    "        key=str(mem_id),\n",
+    "        value={\"text\": content},\n",
+    "    )\n",
+    "    return f\"Stored memory {mem_id}\"\n",
+    "\n",
+    "\n",
+    "agent = create_react_agent(\n",
+    "    init_chat_model(\"openai:gpt-4o-mini\"),\n",
+    "    tools=[upsert_memory],\n",
+    "    # The state_modifier is run to prepare the messages for the LLM. It is called\n",
+    "    # right before each LLM call\n",
+    "    state_modifier=prepare_messages,\n",
+    "    store=store,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "What are you in the mood for? Since you love Italian food and pizza, maybe something in that realm would be great! Would you like suggestions for a specific dish or restaurant?"
+     ]
+    }
+   ],
+   "source": [
+    "for message, metadata in agent.stream(\n",
+    "    input={\"messages\": [{\"role\": \"user\", \"content\": \"I'm hungry\"}]},\n",
+    "    stream_mode=\"messages\",\n",
+    "):\n",
+    "    print(message.content, end=\"\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Advanced Usage\n",
+    "\n",
+    "#### Multi-vector indexing\n",
+    "\n",
+    "Store and search different aspects of memories separately to improve recall or omit certain fields from being indexed."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Expect mem 2\n",
+      "Item: mem2; Score (0.5895009051396596)\n",
+      "Memory: Ate alone at home\n",
+      "Emotion: felt a bit lonely\n",
+      "\n",
+      "Expect mem1\n",
+      "Item: mem1; Score (0.6207546534134083)\n",
+      "Memory: Had pizza with friends at Mario's\n",
+      "Emotion: felt happy and connected\n",
+      "\n",
+      "Expect random lower score (ravioli not indexed)\n",
+      "Item: mem1; Score (0.2686278787315685)\n",
+      "Memory: Had pizza with friends at Mario's\n",
+      "Emotion: felt happy and connected\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Configure store to embed both memory content and emotional context\n",
+    "store = InMemoryStore(\n",
+    "    index={\"embed\": embeddings, \"dims\": 1536, \"fields\": [\"memory\", \"emotional_context\"]}\n",
+    ")\n",
+    "# Store memories with different content/emotion pairs\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem1\",\n",
+    "    {\n",
+    "        \"memory\": \"Had pizza with friends at Mario's\",\n",
+    "        \"emotional_context\": \"felt happy and connected\",\n",
+    "        \"this_isnt_indexed\": \"I prefer ravioli though\",\n",
+    "    },\n",
+    ")\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem2\",\n",
+    "    {\n",
+    "        \"memory\": \"Ate alone at home\",\n",
+    "        \"emotional_context\": \"felt a bit lonely\",\n",
+    "        \"this_isnt_indexed\": \"I like pie\",\n",
+    "    },\n",
+    ")\n",
+    "\n",
+    "# Search focusing on emotional state - matches mem2\n",
+    "results = store.search(\n",
+    "    (\"user_123\", \"memories\"), query=\"times they felt isolated\", limit=1\n",
+    ")\n",
+    "print(\"Expect mem 2\")\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Emotion: {r.value['emotional_context']}\\n\")\n",
+    "\n",
+    "# Search focusing on social eating - matches mem1\n",
+    "print(\"Expect mem1\")\n",
+    "results = store.search((\"user_123\", \"memories\"), query=\"fun pizza\", limit=1)\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Emotion: {r.value['emotional_context']}\\n\")\n",
+    "\n",
+    "print(\"Expect random lower score (ravioli not indexed)\")\n",
+    "results = store.search((\"user_123\", \"memories\"), query=\"ravioli\", limit=1)\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Emotion: {r.value['emotional_context']}\\n\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### Override fields at storage time\n",
+    "You can override which fields to embed when storing a specific memory using `put(..., index=[...fields])`, regardless of the store's default configuration."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Expect mem1\n",
+      "Item: mem1; Score (0.3374968677940555)\n",
+      "Memory: I love spicy food\n",
+      "Context: At a Thai restaurant\n",
+      "\n",
+      "Expect mem2\n",
+      "Item: mem2; Score (0.36784461593247436)\n",
+      "Memory: The restaurant was too loud\n",
+      "Context: Dinner at an Italian place\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "store = InMemoryStore(\n",
+    "    index={\n",
+    "        \"embed\": embeddings,\n",
+    "        \"dims\": 1536,\n",
+    "        \"fields\": [\"memory\"],\n",
+    "    }  # Default to embed memory field\n",
+    ")\n",
+    "\n",
+    "# Store one memory with default indexing\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem1\",\n",
+    "    {\"memory\": \"I love spicy food\", \"context\": \"At a Thai restaurant\"},\n",
+    ")\n",
+    "\n",
+    "# Store another overriding which fields to embed\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem2\",\n",
+    "    {\"memory\": \"The restaurant was too loud\", \"context\": \"Dinner at an Italian place\"},\n",
+    "    index=[\"context\"],  # Override: only embed the context\n",
+    ")\n",
+    "\n",
+    "# Search about food - matches mem1 (using default field)\n",
+    "print(\"Expect mem1\")\n",
+    "results = store.search(\n",
+    "    (\"user_123\", \"memories\"), query=\"what food do they like\", limit=1\n",
+    ")\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Context: {r.value['context']}\\n\")\n",
+    "\n",
+    "# Search about restaurant atmosphere - matches mem2 (using overridden field)\n",
+    "print(\"Expect mem2\")\n",
+    "results = store.search(\n",
+    "    (\"user_123\", \"memories\"), query=\"restaurant environment\", limit=1\n",
+    ")\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Context: {r.value['context']}\\n\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### Disable Indexing for Specific Memories\n",
+    "\n",
+    "Some memories shouldn't be searchable by content. You can disable indexing for these while still storing them using \n",
+    "`put(..., index=False)`. Example:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Expect mem1\n",
+      "Item: mem1; Score (0.32269984224327286)\n",
+      "Memory: I love chocolate ice cream\n",
+      "Type: preference\n",
+      "\n",
+      "Expect low score (mem2 not indexed)\n",
+      "Item: mem1; Score (0.010241633698527089)\n",
+      "Memory: I love chocolate ice cream\n",
+      "Type: preference\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "store = InMemoryStore(index={\"embed\": embeddings, \"dims\": 1536, \"fields\": [\"memory\"]})\n",
+    "\n",
+    "# Store a normal indexed memory\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem1\",\n",
+    "    {\"memory\": \"I love chocolate ice cream\", \"type\": \"preference\"},\n",
+    ")\n",
+    "\n",
+    "# Store a system memory without indexing\n",
+    "store.put(\n",
+    "    (\"user_123\", \"memories\"),\n",
+    "    \"mem2\",\n",
+    "    {\"memory\": \"User completed onboarding\", \"type\": \"system\"},\n",
+    "    index=False,  # Disable indexing entirely\n",
+    ")\n",
+    "\n",
+    "# Search about food preferences - finds mem1\n",
+    "print(\"Expect mem1\")\n",
+    "results = store.search((\"user_123\", \"memories\"), query=\"what food preferences\", limit=1)\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Type: {r.value['type']}\\n\")\n",
+    "\n",
+    "# Search about onboarding - won't find mem2 (not indexed)\n",
+    "print(\"Expect low score (mem2 not indexed)\")\n",
+    "results = store.search((\"user_123\", \"memories\"), query=\"onboarding status\", limit=1)\n",
+    "for r in results:\n",
+    "    print(f\"Item: {r.key}; Score ({r.score})\")\n",
+    "    print(f\"Memory: {r.value['memory']}\")\n",
+    "    print(f\"Type: {r.value['type']}\\n\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.2"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
@@ -0,0 +1,383 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "7c58c957-83d8-44ff-8580-a9b3dd39a0a9",
+   "metadata": {},
+   "source": [
+    "# How to update graph state from tools"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "95f30587-8dd2-40be-920d-59539089c09f",
+   "metadata": {},
+   "source": [
+    "!!! info \"Prerequisites\"\n",
+    "    This guide assumes familiarity with the following:\n",
+    "    \n",
+    "    - [Command](../../concepts/low_level/#command)\n",
+    "\n",
+    "A common use case is updating graph state from inside a tool. For example, in a customer support application you might want to look up customer account number or ID in the beginning of the conversation. To update the graph state from the tool, you can return `Command(update={\"my_custom_key\": \"foo\", \"messages\": [...]})` from the tool:\n",
+    "\n",
+    "```python\n",
+    "@tool\n",
+    "def lookup_user_info(tool_call_id: Annotated[str, InjectedToolCallId], config: RunnableConfig):\n",
+    "    \"\"\"Use this to look up user information to better assist them with their questions.\"\"\"\n",
+    "    user_info = get_user_info(config)\n",
+    "    return Command(\n",
+    "        update={\n",
+    "            # update the state keys\n",
+    "            \"user_info\": user_info,\n",
+    "            # update the message history\n",
+    "            \"messages\": [ToolMessage(\"Successfully looked up user information\", tool_call_id=tool_call_id)]\n",
+    "        }\n",
+    "    )\n",
+    "```\n",
+    "\n",
+    "!!! important\n",
+    "\n",
+    "    If you want to use tools that return `Command` and update graph state, you can either use prebuilt [`create_react_agent`][langgraph.prebuilt.chat_agent_executor.create_react_agent] / [`ToolNode`][langgraph.prebuilt.tool_node.ToolNode] components, or implement your own tool-executing node that collects `Command` objects returned by the tools and returns a list of them, e.g.:\n",
+    "    \n",
+    "    ```python\n",
+    "    def call_tools(state):\n",
+    "        ...\n",
+    "        commands = [tools_by_name[call[\"name\"].invoke(call, config={\"coerce_tool_content\": False}) for tool_call in tool_calls]\n",
+    "        return commands\n",
+    "    ```\n",
+    "\n",
+    "This guide shows how you can do this using LangGraph's prebuilt components ([`create_react_agent`][langgraph.prebuilt.chat_agent_executor.create_react_agent] / [`ToolNode`][langgraph.prebuilt.tool_node.ToolNode]).\n",
+    "\n",
+    "!!! note\n",
+    "\n",
+    "    Support for tools that return [`Command`][langgraph.types.Command] was added in LangGraph `v0.2.57`.\n",
+    "\n",
+    "## Setup\n",
+    "\n",
+    "First, let's install the required packages and set our API keys:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "64500eca-1cdc-43d9-9401-f4cd9999881f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%%capture --no-stderr\n",
+    "%pip install -U langgraph langchain-openai"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "a3f92fb2-9175-47fa-9c7d-ad5f44bfd20e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Please provide your OPENAI_API_KEY ········\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import getpass\n",
+    "\n",
+    "\n",
+    "def _set_if_undefined(var: str):\n",
+    "    if not os.environ.get(var):\n",
+    "        os.environ[var] = getpass.getpass(f\"Please provide your {var}\")\n",
+    "\n",
+    "\n",
+    "_set_if_undefined(\"OPENAI_API_KEY\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "caf6ff9f-c1e6-499e-a230-9fa231ea7d2f",
+   "metadata": {},
+   "source": [
+    "<div class=\"admonition tip\">\n",
+    "    <p class=\"admonition-title\">Set up <a href=\"https://smith.langchain.com\">LangSmith</a> for LangGraph development</p>\n",
+    "    <p style=\"padding-top: 5px;\">\n",
+    "        Sign up for LangSmith to quickly spot issues and improve the performance of your LangGraph projects. LangSmith lets you use trace data to debug, test, and monitor your LLM apps built with LangGraph — read more about how to get started <a href=\"https://docs.smith.langchain.com\">here</a>. \n",
+    "    </p>\n",
+    "</div>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "10e9a9c6-fa3f-416c-bac0-3e58d7259908",
+   "metadata": {},
+   "source": [
+    "Let's create a simple ReAct style agent that can look up user information and personalize the response based on the user info."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4255b9b9-cf67-4cc3-8018-1708f5dfcfd2",
+   "metadata": {},
+   "source": [
+    "## Define tool"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7de6b010-aab1-4fe8-8251-907fcae78583",
+   "metadata": {},
+   "source": [
+    "First, let's define the tool that we'll be using to look up user information. We'll use a naive implementation that simply looks user information up using a dictionary:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8d070c9f-6e61-4724-85dc-ac4531b9c79a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "USER_INFO = [\n",
+    "    {\"user_id\": \"1\", \"name\": \"Bob Dylan\", \"location\": \"New York, NY\"},\n",
+    "    {\"user_id\": \"2\", \"name\": \"Taylor Swift\", \"location\": \"Beverly Hills, CA\"},\n",
+    "]\n",
+    "\n",
+    "USER_ID_TO_USER_INFO = {info[\"user_id\"]: info for info in USER_INFO}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "08d1ecca-ee57-4e97-b8d0-e09de85337d4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langgraph.prebuilt.chat_agent_executor import AgentState\n",
+    "from langgraph.types import Command\n",
+    "from langchain_core.tools import tool\n",
+    "from langchain_core.tools.base import InjectedToolCallId\n",
+    "from langchain_core.messages import ToolMessage\n",
+    "from langchain_core.runnables import RunnableConfig\n",
+    "\n",
+    "from typing_extensions import Any, Annotated\n",
+    "\n",
+    "\n",
+    "class State(AgentState):\n",
+    "    # user provided\n",
+    "    last_name: str\n",
+    "    # updated by the tool\n",
+    "    user_info: dict[str, Any]\n",
+    "\n",
+    "\n",
+    "@tool\n",
+    "def lookup_user_info(\n",
+    "    tool_call_id: Annotated[str, InjectedToolCallId], config: RunnableConfig\n",
+    "):\n",
+    "    \"\"\"Use this to look up user information to better assist them with their questions.\"\"\"\n",
+    "    user_id = config.get(\"configurable\", {}).get(\"user_id\")\n",
+    "    if user_id is None:\n",
+    "        raise ValueError(\"Please provide user ID\")\n",
+    "\n",
+    "    if user_id not in USER_ID_TO_USER_INFO:\n",
+    "        raise ValueError(f\"User '{user_id}' not found\")\n",
+    "\n",
+    "    user_info = USER_ID_TO_USER_INFO[user_id]\n",
+    "    return Command(\n",
+    "        update={\n",
+    "            # update the state keys\n",
+    "            \"user_info\": user_info,\n",
+    "            # update the message history\n",
+    "            \"messages\": [\n",
+    "                ToolMessage(\n",
+    "                    \"Successfully looked up user information\", tool_call_id=tool_call_id\n",
+    "                )\n",
+    "            ],\n",
+    "        }\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b99e5f24-5e5e-4a34-baae-467182675bb5",
+   "metadata": {},
+   "source": [
+    "## Define prompt"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cbb06aea-6654-4245-91f8-af6e8f2b5377",
+   "metadata": {},
+   "source": [
+    "Let's now add personalization: we'll respond differently to the user based on the state values AFTER the state has been updated from the tool. To achieve this, let's define a function that will dynamically construct the system prompt based on the graph state. It will be called ever time the LLM is called and the function output will be passed to the LLM:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "c553d062-d145-4145-84bd-9b798f7c95c2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def state_modifier(state: State):\n",
+    "    user_info = state.get(\"user_info\")\n",
+    "    if user_info is None:\n",
+    "        return state[\"messages\"]\n",
+    "\n",
+    "    system_msg = (\n",
+    "        f\"User name is {user_info['name']}. User lives in {user_info['location']}\"\n",
+    "    )\n",
+    "    return [{\"role\": \"system\", \"content\": system_msg}] + state[\"messages\"]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c5acdd5d-68be-466b-9c21-46cbed91d2bc",
+   "metadata": {},
+   "source": [
+    "## Define graph"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "afb65028-0359-46c8-b09c-ffc90180f759",
+   "metadata": {},
+   "source": [
+    "Finally, let's combine this into a single graph using the prebuilt `create_react_agent`:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "2d59db29-fd51-4d29-9854-21763a4855e3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langgraph.prebuilt import create_react_agent\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "\n",
+    "model = ChatOpenAI(model=\"gpt-4o\")\n",
+    "\n",
+    "agent = create_react_agent(\n",
+    "    model,\n",
+    "    # pass the tool that can update state\n",
+    "    [lookup_user_info],\n",
+    "    state_schema=State,\n",
+    "    # pass dynamic prompt function\n",
+    "    state_modifier=state_modifier,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0782b8ab-a603-47b8-9a76-77f593402678",
+   "metadata": {},
+   "source": [
+    "## Use it!"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6165e153-ab28-4404-adea-796c7bd0701b",
+   "metadata": {},
+   "source": [
+    "Let's now try running our agent. We'll need to provide user ID in the config so that our tool knows what information to look up:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "de34a58b-1765-4b63-a232-d46790aff884",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'agent': {'messages': [AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_7LSUh6ZDvGJAUvlWvXiCK4Gf', 'function': {'arguments': '{}', 'name': 'lookup_user_info'}, 'type': 'function'}], 'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 11, 'prompt_tokens': 56, 'total_tokens': 67, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-08-06', 'system_fingerprint': 'fp_9d50cd990b', 'finish_reason': 'tool_calls', 'logprobs': None}, id='run-57eeb216-e35d-4501-aaac-b5c6b26fb17c-0', tool_calls=[{'name': 'lookup_user_info', 'args': {}, 'id': 'call_7LSUh6ZDvGJAUvlWvXiCK4Gf', 'type': 'tool_call'}], usage_metadata={'input_tokens': 56, 'output_tokens': 11, 'total_tokens': 67, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]}}\n",
+      "\n",
+      "\n",
+      "{'tools': {'user_info': {'user_id': '1', 'name': 'Bob Dylan', 'location': 'New York, NY'}, 'messages': [ToolMessage(content='Successfully looked up user information', name='lookup_user_info', id='168d8ff8-b021-4c8b-a11a-3b50c30a072c', tool_call_id='call_7LSUh6ZDvGJAUvlWvXiCK4Gf')]}}\n",
+      "\n",
+      "\n",
+      "{'agent': {'messages': [AIMessage(content=\"Hi Bob! Since you're in New York, NY, there are plenty of exciting things to do over the weekend. Here are some suggestions:\\n\\n1. **Explore Central Park**: Take a leisurely walk, rent a bike, or have a picnic in this iconic park.\\n\\n2. **Visit a Museum**: Check out The Metropolitan Museum of Art or the Museum of Modern Art (MoMA) for an enriching cultural experience.\\n\\n3. **Broadway Show**: Catch a Broadway show or an off-Broadway performance for some world-class entertainment.\\n\\n4. **Food Tour**: Explore different neighborhoods like Greenwich Village or Williamsburg for diverse culinary experiences.\\n\\n5. **Brooklyn Bridge Walk**: Take a walk across the Brooklyn Bridge for stunning views of the city skyline.\\n\\n6. **Visit a Rooftop Bar**: Enjoy a drink with a view at one of New York’s many rooftop bars.\\n\\n7. **Explore a New Neighborhood**: Discover the unique charm of areas like SoHo, Chelsea, or Astoria.\\n\\n8. **Live Music**: Check out live music venues for a night of great performances.\\n\\n9. **Art Galleries**: Visit some of the smaller art galleries around Chelsea or the Lower East Side.\\n\\n10. **Attend a Local Event**: Look up any local events or festivals happening this weekend.\\n\\nFeel free to let me know if you want more details on any of these activities!\", additional_kwargs={'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 285, 'prompt_tokens': 95, 'total_tokens': 380, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-08-06', 'system_fingerprint': 'fp_9d50cd990b', 'finish_reason': 'stop', 'logprobs': None}, id='run-f13ce15b-02b6-40e6-8264-c4d9edd0d03a-0', usage_metadata={'input_tokens': 95, 'output_tokens': 285, 'total_tokens': 380, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]}}\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "for chunk in agent.stream(\n",
+    "    {\"messages\": [(\"user\", \"hi, what should i do this weekend?\")]},\n",
+    "    # provide user ID in the config\n",
+    "    {\"configurable\": {\"user_id\": \"1\"}},\n",
+    "):\n",
+    "    print(chunk)\n",
+    "    print(\"\\n\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d9b2281f-269c-41dd-b6b2-4c743f11ffc9",
+   "metadata": {},
+   "source": [
+    "We can see that the model correctly recommended some New York activities for Bob Dylan! Let's try getting recommendations for Taylor Swift:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "9d71af94-572a-4961-88a7-665e792cf96a",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'agent': {'messages': [AIMessage(content='', additional_kwargs={'tool_calls': [{'id': 'call_5HLtJtzcgmKbtmK6By21wW5Y', 'function': {'arguments': '{}', 'name': 'lookup_user_info'}, 'type': 'function'}], 'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 11, 'prompt_tokens': 56, 'total_tokens': 67, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-08-06', 'system_fingerprint': 'fp_c7ca0ebaca', 'finish_reason': 'tool_calls', 'logprobs': None}, id='run-bacacd7d-76cc-4f6b-9e9b-d9e6f00b9391-0', tool_calls=[{'name': 'lookup_user_info', 'args': {}, 'id': 'call_5HLtJtzcgmKbtmK6By21wW5Y', 'type': 'tool_call'}], usage_metadata={'input_tokens': 56, 'output_tokens': 11, 'total_tokens': 67, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]}}\n",
+      "\n",
+      "\n",
+      "{'tools': {'user_info': {'user_id': '2', 'name': 'Taylor Swift', 'location': 'Beverly Hills, CA'}, 'messages': [ToolMessage(content='Successfully looked up user information', name='lookup_user_info', id='d81ef31e-6d77-4f13-ae86-e2e6ba567e3d', tool_call_id='call_5HLtJtzcgmKbtmK6By21wW5Y')]}}\n",
+      "\n",
+      "\n",
+      "{'agent': {'messages': [AIMessage(content=\"Hi Taylor! Since you're in Beverly Hills, here are a few suggestions for a fun weekend:\\n\\n1. **Hiking at Runyon Canyon**: Enjoy a scenic hike with beautiful views of Los Angeles. It's a great way to get some exercise and enjoy the outdoors.\\n\\n2. **Visit Rodeo Drive**: Spend some time shopping or window shopping at the famous Rodeo Drive. You might even spot some celebrities!\\n\\n3. **Explore the Getty Center**: Check out the art collections and beautiful gardens at the Getty Center. The architecture and views are stunning.\\n\\n4. **Relax at a Spa**: Treat yourself to a relaxing day at one of Beverly Hills' luxurious spas.\\n\\n5. **Dining Out**: Try a new restaurant or visit your favorite spot for a delicious meal. Beverly Hills has a fantastic dining scene.\\n\\n6. **Attend a Local Event**: Check out any local events or concerts happening this weekend. Beverly Hills often hosts exciting events.\\n\\nEnjoy your weekend!\", additional_kwargs={'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 198, 'prompt_tokens': 95, 'total_tokens': 293, 'completion_tokens_details': {'accepted_prediction_tokens': 0, 'audio_tokens': 0, 'reasoning_tokens': 0, 'rejected_prediction_tokens': 0}, 'prompt_tokens_details': {'audio_tokens': 0, 'cached_tokens': 0}}, 'model_name': 'gpt-4o-2024-08-06', 'system_fingerprint': 'fp_c7ca0ebaca', 'finish_reason': 'stop', 'logprobs': None}, id='run-2057df76-f192-4c69-a66a-1f0a86bf5d66-0', usage_metadata={'input_tokens': 95, 'output_tokens': 198, 'total_tokens': 293, 'input_token_details': {'audio': 0, 'cache_read': 0}, 'output_token_details': {'audio': 0, 'reasoning': 0}})]}}\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "for chunk in agent.stream(\n",
+    "    {\"messages\": [(\"user\", \"hi, what should i do this weekend?\")]},\n",
+    "    {\"configurable\": {\"user_id\": \"2\"}},\n",
+    "):\n",
+    "    print(chunk)\n",
+    "    print(\"\\n\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
@@ -13,3 +13,5 @@
        - PregelExecutableTask
        - StateSnapshot
        - Send
+        - Command
+        - interrupt
@@ -6,25 +6,24 @@ title: Tutorials

 # Tutorials

-Welcome to the LangGraph Tutorials! These notebooks introduce LangGraph through building various language agents and applications.
+New to LangGraph or LLM app development? Read this material to get up and running building your first applications.

-## Quick Start
+## Get Started 🚀 {#quick-start}

-Learn the basics of LangGraph through a comprehensive quick start in which you will build an agent from scratch.
+- [LangGraph Quickstart](introduction.ipynb): Build a chatbot that can use tools and keep track of conversation history. Add human-in-the-loop capabilities and explore how time-travel works.
+- [LangGraph Server Quickstart](langgraph-platform/local-server.md): Launch a LangGraph server locally and interact with it using the REST API and LangGraph Studio Web UI.
+- [LangGraph Cloud QuickStart](../cloud/quick_start.md): Deploy a LangGraph app using LangGraph Cloud.
+- [LangGraph Template Quickstart](../concepts/template_applications.md): Quickly start building with LangGraph Platform using a template application.

- [Quick Start](introduction.ipynb): In this tutorial, you will build a support chatbot using LangGraph.
- [LangGraph Cloud Quick Start](../cloud/quick_start.md): In this tutorial, you will build and deploy an agent to LangGraph Cloud.
+## Use cases 🛠️

-## Use cases
-
-Learn from example implementations of graphs designed for specific scenarios and that implement common design patterns.
+Explore practical implementations tailored for specific scenarios:

 ### Chatbots

- [Customer Support](customer-support/customer-support.ipynb): Build a customer support chatbot to manage flights, hotel reservations, car rentals, and other tasks
- [Prompt Generation from User Requirements](chatbots/information-gather-prompting.ipynb): Build an information gathering chatbot
- [Code Assistant](code_assistant/langgraph_code_assistant.ipynb): Build a code analysis and generation assistant
-
+- [Customer Support](customer-support/customer-support.ipynb): Build a multi-functional support bot for flights, hotels, and car rentals.
+- [Prompt Generation from User Requirements](chatbots/information-gather-prompting.ipynb): Build an information gathering chatbot.
+- [Code Assistant](code_assistant/langgraph_code_assistant.ipynb): Build a code analysis and generation assistant.

 ### RAG

@@ -5,21 +5,21 @@
   "id": "4a1aae78-88a6-4133-b905-7e46c8e3772f",
   "metadata": {},
   "source": [
-    "# LangGraph Quick Start\n",
+    "# 🚀 LangGraph Quick Start\n",
    "\n",
-    "In this comprehensive quick start, we will build a support chatbot in LangGraph that can:\n",
+    "In this tutorial, we will build a support chatbot in LangGraph that can:\n",
    "\n",
-    "- Answer common questions by searching the web\n",
-    "- Maintain conversation state across calls\n",
-    "- Route complex queries to a human for review\n",
-    "- Use custom state to control its behavior\n",
-    "- Rewind and explore alternative conversation paths\n",
+    "✅ **Answer common questions** by searching the web  \n",
+    "✅ **Maintain conversation state** across calls  \n",
+    "✅ **Route complex queries** to a human for review  \n",
+    "✅ **Use custom state** to control its behavior  \n",
+    "✅ **Rewind and explore** alternative conversation paths  \n",
    "\n",
-    "We'll start with a basic chatbot and progressively add more sophisticated capabilities, introducing key LangGraph concepts along the way.\n",
+    "We'll start with a **basic chatbot** and progressively add more sophisticated capabilities, introducing key LangGraph concepts along the way. Let’s dive in! 🌟\n",
    "\n",
    "## Setup\n",
    "\n",
-    "First, install the required packages:"
+    "First, install the required packages and configure your environment:"
   ]
  },
  {
@@ -33,14 +33,6 @@
    "%pip install -U langgraph langsmith langchain_anthropic"
   ]
  },
-  {
-   "cell_type": "markdown",
-   "id": "a6d1e870-1bc0-4d44-86c0-96681ccf6113",
-   "metadata": {},
-   "source": [
-    "Next, set your API keys:"
-   ]
-  },
  {
   "cell_type": "code",
   "execution_count": 2,
@@ -120,27 +112,24 @@
   ]
  },
  {
+   "attachments": {},
   "cell_type": "markdown",
-   "id": "31c755cd-8994-4867-bdff-96a55d7beae7",
+   "id": "c08c41da-0855-49d3-9a3d-b7eb94413367",
   "metadata": {},
   "source": [
-    "<div class=\"admonition tip\">\n",
-    "    <p class=\"admonition-title\">Note</p>\n",
-    "    <p>\n",
-    "    The first thing you do when you define a graph is define the <code>State</code> of the graph. The <code>State</code> consists of the schema of the graph as well as <a href=\"https://langchain-ai.github.io/langgraph/concepts/low_level/#reducers\">reducer functions</a> which specify how to apply updates to the state. In our example <code>State</code> is a <code>TypedDict</code> with a single key: <code>messages</code>. The <code>messages</code> key is annotated with the <a href=\"https://langchain-ai.github.io/langgraph/reference/graphs/#langgraph.graph.message.add_messages\"><code>add_messages</code></a> reducer function, which tells LangGraph to append new messages to the existing list, rather than overwriting it. State keys without an annotation will be overwritten by each update, storing the most recent value. Check out <a href=\"https://langchain-ai.github.io/langgraph/reference/graphs/#langgraph.graph.message.add_messages\">this conceptual guide</a> to learn more about state, reducers and other low-level concepts.\n",
-    "    </p>\n",
-    "</div>"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4137feed-746e-4c72-a34a-f7a699ad5dcf",
-   "metadata": {},
-   "source": [
-    "So now our graph knows two things:\n",
+    "Our graph can now handle two key tasks:\n",
+    "\n",
+    "1. Each `node` can receive the current `State` as input and output an update to the state.\n",
+    "2. Updates to `messages` will be appended to the existing list rather than overwriting it, thanks to the prebuilt [`add_messages`](https://langchain-ai.github.io/langgraph/reference/graphs/?h=add+messages#add_messages) function used with the `Annotated` syntax.\n",
+    "\n",
+    "------\n",
+    "\n",
+    "!!! tip \"Concept\"\n",
+    "\n",
+    "    When defining a graph, the first step is to define its `State`. The `State` includes the graph's schema and [reducer functions](https://langchain-ai.github.io/langgraph/concepts/low_level/#reducers) that handle state updates. In our example, `State` is a `TypedDict` with one key: `messages`. The [`add_messages`](https://langchain-ai.github.io/langgraph/reference/graphs/#langgraph.graph.message.add_messages) reducer function is used to append new messages to the list instead of overwriting it. Keys without a reducer annotation will overwrite previous values. Learn more about state, reducers, and related concepts in [this guide](https://langchain-ai.github.io/langgraph/reference/graphs/#langgraph.graph.message.add_messages).\n",
+    "\n",
+    "---------\n",
    "\n",
-    "1. Every `node` we define will receive the current `State` as input and return a value that updates that state.\n",
-    "2. `messages` will be _appended_ to the current list, rather than directly overwritten. This is communicated via the prebuilt [`add_messages`](https://langchain-ai.github.io/langgraph/reference/graphs/?h=add+messages#add_messages) function in the `Annotated` syntax.\n",
    "\n",
    "Next, add a \"`chatbot`\" node. Nodes represent units of work. They are typically regular python functions."
   ]
@@ -365,7 +354,7 @@
   "id": "f22c5d4a-3134-413c-81fe-dd9752fbeb66",
   "metadata": {},
   "source": [
-    "## Part 2: Enhancing the Chatbot with Tools\n",
+    "## Part 2: 🛠️ Enhancing the Chatbot with Tools\n",
    "\n",
    "To handle queries our chatbot can't answer \"from memory\", we'll integrate a web search tool. Our bot can use this tool to find relevant information and provide better responses.\n",
    "\n",
@@ -2046,7 +2035,7 @@
    "\n",
    "So far, we've relied on a simple state (it's just a list of messages!). You can go far with this simple state, but if you want to define complex behavior without relying on the message list, you can add additional fields to the state. In this section, we will extend our chat bot with a new node to illustrate this.\n",
    "\n",
-    "In the examples above, we involved a human deterministically: the graph __always__ interrupted whenever an tool was invoked. Suppose we wanted our chat bot to have the choice of relying on a human.\n",
+    "In the examples above, we involved a human deterministically: the graph __always__ interrupted whenever a tool was invoked. Suppose we wanted our chat bot to have the choice of relying on a human.\n",
    "\n",
    "One way to do this is to create a passthrough \"human\" node, before which the graph will always stop. We will only execute this node if the LLM invokes a \"human\" tool. For our convenience, we will include an \"ask_human\" flag in our graph state that we will flip if the LLM calls this tool.\n",
    "\n",
@@ -3136,11 +3125,29 @@
   "id": "e584d57f-5aad-4507-815f-0b2e4b64b791",
   "metadata": {},
   "source": [
-    "## Conclusion\n",
+    "## Next Steps\n",
    "\n",
-    "Congrats! You've completed the intro tutorial and built a chat bot in LangGraph that supports tool calling, persistent memory, human-in-the-loop interactivity, and even time-travel!\n",
+    "Take your journey further by exploring deployment and advanced features:\n",
    "\n",
-    "The [LangGraph documentation](https://langchain-ai.github.io/langgraph/) is a great resource for diving deeper into the library's capabilities."
+    "### Server Quickstart\n",
+    "\n",
+    "- **[LangGraph Server Quickstart](../langgraph-platform/local-server)**: Launch a LangGraph server locally and interact with it using the REST API and LangGraph Studio Web UI.\n",
+    "\n",
+    "### LangGraph Cloud\n",
+    "\n",
+    "- **[LangGraph Cloud QuickStart](../../cloud/quick_start)**: Deploy your LangGraph app using LangGraph Cloud.\n",
+    "\n",
+    "### LangGraph Framework\n",
+    "\n",
+    "- **[LangGraph Concepts](../../concepts)**: Learn the foundational concepts of LangGraph.  \n",
+    "- **[LangGraph How-to Guides](../../how-tos)**: Guides for common tasks with LangGraph.\n",
+    "\n",
+    "### LangGraph Platform\n",
+    "\n",
+    "Expand your knowledge with these resources:\n",
+    "\n",
+    "- **[LangGraph Platform Concepts](../../concepts#langgraph-platform)**: Understand the foundational concepts of the LangGraph Platform.  \n",
+    "- **[LangGraph Platform How-to Guides](../../how-tos#langgraph-platform)**: Guides for common tasks with LangGraph Platform. "
   ]
  }
 ],
@@ -3160,7 +3167,7 @@
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
+   "version": "3.11.4"
  }
 },
 "nbformat": 4,
@@ -0,0 +1,253 @@
+# Quick Start: Launch Local LangGraph Server
+
+This is a quick start guide to help you get a LangGraph app up and running locally.
+
+!!! info "Requirements"
+
+    - Python >= 3.11
+    - [LangGraph CLI](https://langchain-ai.github.io/langgraph/cloud/reference/cli/): Requires langchain-cli[inmem] >= 0.1.58
+
+## Install the LangGraph CLI
+
+```bash
+pip install -U "langgraph-cli[inmem]" python-dotenv
+```
+
+## 🌱 Create a LangGraph App
+
+Create a new app from the `react-agent` template. This template is a simple agent that can be flexibly extended to many tools.
+
+=== "Python Server"
+
+    ```shell
+    langgraph new path/to/your/app --template react-agent-python 
+    ```
+
+=== "Node Server"
+
+    ```shell
+    langgraph new path/to/your/app --template react-agent-js
+    ```
+
+!!! tip "Additional Templates"
+
+    If you use `langgraph new` without specifying a template, you will be presented with an interactive menu that will allow you to choose from a list of available templates.
+
+## Install Dependencies
+
+In the root of your new LangGraph app, install the dependencies in `edit` mode so your local changes are used by the server:
+
+```shell
+pip install -e .
+```
+
+## Create a `.env` file
+
+You will find a `.env.example` in the root of your new LangGraph app. Create
+a `.env` file in the root of your new LangGraph app and copy the contents of the `.env.example` file into it, filling in the necessary API keys:
+
+```bash
+LANGSMITH_API_KEY=lsv2...
+TAVILY_API_KEY=tvly-...
+ANTHROPIC_API_KEY=sk-
+OPENAI_API_KEY=sk-...
+```
+
+<details><summary>Get API Keys</summary>
+    <ul>
+        <li> <b>LANGSMITH_API_KEY</b>: Go to the <a href="https://smith.langchain.com/settings">LangSmith Settings page</a>. Then clck <b>Create API Key</b>.
+        </li>
+        <li>
+            <b>ANTHROPIC_API_KEY</b>: Get an API key from <a href="https://console.anthropic.com/">Anthropic</a>.
+        </li>
+        <li>
+            <b>OPENAI_API_KEY</b>: Get an API key from <a href="https://openai.com/">OpenAI</a>.
+        </li>
+        <li>
+            <b>TAVILY_API_KEY</b>: Get an API key on the <a href="https://app.tavily.com/">Tavily website</a>.
+        </li>
+    </ul>
+</details>
+
+## 🚀 Launch LangGraph Server
+
+```shell
+langgraph dev
+```
+
+This will start up the LangGraph API server locally. If this runs successfully, you should see something like:
+
+>    Ready!
+> 
+>    - API: [http://localhost:8123](http://localhost:8123/)
+>     
+>    - Docs: http://localhost:8123/docs
+>     
+>    - LangGraph Studio Web UI: https://smith.langchain.com/studio/?baseUrl=http://127.0.0.1:8123
+
+
+!!! note "In-Memory Mode"
+
+    The `langgraph dev` command starts LangGraph Server in an in-memory mode. This mode is suitable for development and testing purposes. For production use, you should deploy LangGraph Server with access to a persistent storage backend.
+
+    If you want to test your application with a persistent storage backend, you can use the `langgraph up` command instead of `langgraph dev`. You will
+    need to have `docker` installed on your machine to use this command.
+
+## LangGraph Studio Web UI
+
+Test your graph in the LangGraph Studio Web UI by visiting the URL provided in the output of the `langgraph up` command.
+
+>    - LangGraph Studio Web UI: https://smith.langchain.com/studio/?baseUrl=http://127.0.0.1:8123
+
+!!! warning "Safari Compatibility"
+    
+    Currently, LangGraph Studio Web does not support Safari when running a server locally.
+
+## Test the API
+
+=== "Python SDK (Async)"
+
+    **Install the LangGraph Python SDK**
+
+    ```shell
+    pip install langgraph-sdk
+    ```
+
+    **Send a message to the assistant (threadless run)**
+
+    ```python
+    from langgraph_sdk import get_client
+
+    client = get_client(url="http://localhost:8123")
+
+    async for chunk in client.runs.stream(
+        None,  # Threadless run
+        "agent", # Name of assistant. Defined in langgraph.json.
+        input={
+            "messages": [{
+                "role": "human",
+                "content": "What is LangGraph?",
+            }],
+        },
+        stream_mode="updates",
+    ):
+        print(f"Receiving new event of type: {chunk.event}...")
+        print(chunk.data)
+        print("\n\n")
+    ```
+
+=== "Python SDK (Sync)"
+
+    **Install the LangGraph Python SDK**
+
+    ```shell
+    pip install langgraph-sdk
+    ```
+
+    **Send a message to the assistant (threadless run)**
+
+    ```python
+    from langgraph_sdk import get_sync_client
+
+    client = get_sync_client(url="http://localhost:8123")
+
+    for chunk in client.runs.stream(
+        None,  # Threadless run
+        "agent", # Name of assistant. Defined in langgraph.json.
+        input={
+            "messages": [{
+                "role": "human",
+                "content": "What is LangGraph?",
+            }],
+        },
+        stream_mode="updates",
+    ):
+        print(f"Receiving new event of type: {chunk.event}...")
+        print(chunk.data)
+        print("\n\n")
+    ```
+
+=== "Javascript SDK"
+
+    **Install the LangGraph JS SDK**
+
+    ```shell
+    npm install @langchain/langgraph-sdk
+    ```
+
+    **Send a message to the assistant (threadless run)**
+
+    ```js
+    const { Client } = await import("@langchain/langgraph-sdk");
+
+    // only set the apiUrl if you changed the default port when calling langgraph up
+    const client = new Client({ apiUrl: "http://localhost:8123"});
+
+    const streamResponse = client.runs.stream(
+        null, // Threadless run
+        "agent", // Assistant ID
+        {
+            input: {
+                "messages": [
+                    { "role": "user", "content": "What is LangGraph?"}
+                ]
+            },
+            streamMode: "messages",
+        }
+    );
+
+    for await (const chunk of streamResponse) {
+        console.log(`Receiving new event of type: ${chunk.event}...`);
+        console.log(JSON.stringify(chunk.data));
+        console.log("\n\n");
+    }
+    ```
+
+=== "Rest API"
+
+    ```bash
+    curl -s --request POST \
+        --url "http://localhost:8123/runs/stream" \
+        --header 'Content-Type: application/json' \
+        --data "{
+            \"assistant_id\": \"agent\",
+            \"input\": {
+                \"messages\": [
+                    {
+                        \"role\": \"human\",
+                        \"content\": \"What is LangGraph?\"
+                    }
+                ]
+            },
+            \"stream_mode\": \"updates\"
+        }" 
+    ```
+
+!!! tip "Auth"
+
+    If you're connecting to a remote server, you will need to provide a LangSmith
+    API Key for authorization. Please see the API Reference for the clients
+    for more information.
+
+## Next Steps
+
+Now that you have a LangGraph app running locally, take your journey further by exploring deployment and advanced features:
+
+### 🌐 Deploy to LangGraph Cloud
+
+- **[LangGraph Cloud QuickStart](../../cloud/quick_start.md)**: Deploy your LangGraph app using LangGraph Cloud.
+
+### 📚 Learn More about LangGraph Platform
+
+Expand your knowledge with these resources:
+
+- **[LangGraph Platform Concepts](../../concepts/index.md#langgraph-platform)**: Understand the foundational concepts of the LangGraph Platform.  
+- **[LangGraph Platform How-to Guides](../../how-tos/index.md#langgraph-platform)**: Discover step-by-step guides to build and deploy applications.
+
+### 🛠️ Developer References
+
+Access detailed documentation for development and API usage:
+
+- **[LangGraph Server API Reference](../../cloud/reference/api/api_ref.html)**: Explore the LangGraph Server API documentation.  
+- **[Python SDK Reference](../../cloud/reference/sdk/python_sdk_ref.md)**: Explore the Python SDK API Reference.
+- **[JS/TS SDK Reference](../../cloud/reference/sdk/js_ts_sdk_ref.md)**: Explore the Python SDK API Reference.
@@ -934,7 +934,7 @@
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
+   "version": "3.11.4"
  }
 },
 "nbformat": 4,
@@ -112,7 +112,7 @@
   "metadata": {},
   "outputs": [],
   "source": [
-    "_set_env(\"LANGCHAIN_API_KEY\")\n",
+    "_set_env(\"LANGSMITH_API_KEY\")\n",
    "os.environ[\"LANGCHAIN_TRACING_V2\"] = \"true\"\n",
    "os.environ[\"LANGCHAIN_PROJECT\"] = \"local-llama32-rag\""
   ]
@@ -43,7 +43,7 @@
   "outputs": [],
   "source": [
    "%%capture --no-stderr\n",
-    "%pip install -U langgraph langchain_anthropic langsmith\n",
+    "%pip install -U langgraph langchain_anthropic langsmith langchain-community\n",
    "%pip install -U sklearn langchain_openai"
   ]
  },
@@ -632,7 +632,7 @@
   "metadata": {},
   "outputs": [],
   "source": [
-    "from langchain.cache import InMemoryCache\n",
+    "from langchain_community.cache import InMemoryCache\n",
    "from langchain.globals import set_llm_cache\n",
    "\n",
    "# Optional. If you are running into errors or rate limits and want to avoid repeated computation,\n",
@@ -94,6 +94,7 @@ nav:
      - Quick Start:
        - Quick Start: tutorials#quick-start
        - tutorials/introduction.ipynb
+        - tutorials/langgraph-platform/local-server.md
        - cloud/quick_start.md
      - Chatbots:
          - Chatbots: tutorials#chatbots
@@ -150,6 +151,7 @@ nav:
            - how-tos/branching.ipynb
            - how-tos/map-reduce.ipynb
            - how-tos/recursion-limit.ipynb
+            - how-tos/command.ipynb
        - Persistence:
            - Persistence: how-tos#persistence
            - how-tos/persistence.ipynb
@@ -163,6 +165,7 @@ nav:
            - how-tos/memory/manage-conversation-history.ipynb
            - how-tos/memory/delete-messages.ipynb
            - how-tos/memory/add-summary-conversation-history.ipynb
+            - how-tos/memory/semantic-search.ipynb
        - Human-in-the-loop:
            - Human-in-the-loop: how-tos#human-in-the-loop
            - how-tos/human_in_the_loop/breakpoints.ipynb
@@ -189,6 +192,7 @@ nav:
            - how-tos/tool-calling.ipynb
            - how-tos/tool-calling-errors.ipynb
            - how-tos/pass-run-time-values-to-tools.ipynb
+            - how-tos/update-state-from-tools.ipynb
            - how-tos/pass-config-to-tools.ipynb
            - how-tos/many-tools.ipynb
        - Subgraphs:
@@ -196,6 +200,8 @@ nav:
            - how-tos/subgraph.ipynb
            - how-tos/subgraphs-manage-state.ipynb
            - how-tos/subgraph-transform-state.ipynb
+        - Multi-agent:
+            - how-tos/multi-agent-network.ipynb
        - State Management:
            - State Management: how-tos#state-management
            - how-tos/state-model.ipynb
@@ -224,6 +230,7 @@ nav:
              - cloud/deployment/setup.md
              - cloud/deployment/setup_pyproject.md
              - cloud/deployment/setup_javascript.md
+              - cloud/deployment/semantic_search.md
              - cloud/deployment/custom_docker.md
              - cloud/deployment/test_locally.md
              - cloud/deployment/graph_rebuild.md
@@ -438,4 +445,4 @@ validation:
  # and those anchors are not available in the actual doc
  anchors: info
  # this is needed to handle headers with anchors for nav
-  not_found: info
+  not_found: info
@@ -42,7 +42,7 @@ class DuckDBSaver(BaseDuckDBSaver):
            DuckDBSaver: A new DuckDBSaver instance.
        """
        with duckdb.connect(conn_string) as conn:
-            yield DuckDBSaver(conn)
+            yield cls(conn)

    def setup(self) -> None:
        """Set up the checkpoint database asynchronously.
@@ -45,7 +45,7 @@ class AsyncDuckDBSaver(BaseDuckDBSaver):
            AsyncDuckDBSaver: A new AsyncDuckDBSaver instance.
        """
        with duckdb.connect(conn_string) as conn:
-            yield AsyncDuckDBSaver(conn)
+            yield cls(conn)

    async def setup(self) -> None:
        """Set up the checkpoint database asynchronously.
@@ -156,7 +156,7 @@ class AsyncDuckDBStore(AsyncBatchedBaseStore, BaseDuckDBStore):
            AsyncDuckDBStore: A new AsyncDuckDBStore instance.
        """
        with duckdb.connect(conn_string) as conn:
-            yield AsyncDuckDBStore(conn)
+            yield cls(conn)

    async def setup(self) -> None:
        """Set up the store database asynchronously.
@@ -23,6 +23,7 @@ from langgraph.store.base import (
    Op,
    PutOp,
    Result,
+    SearchItem,
    SearchOp,
 )

@@ -283,7 +284,7 @@ class DuckDBStore(BaseStore, BaseDuckDBStore[duckdb.DuckDBPyConnection]):

        for cur, idx in cursors:
            rows = cur.fetchall()
-            items = [_row_to_item(_convert_ns(row[0]), row) for row in rows]
+            items = [_row_to_search_item(_convert_ns(row[0]), row) for row in rows]
            results[idx] = items

    def _batch_list_namespaces_ops(
@@ -376,6 +377,22 @@ def _row_to_item(
    )


+def _row_to_search_item(
+    namespace: tuple[str, ...],
+    row: tuple,
+) -> SearchItem:
+    """Convert a row from the database into an SearchItem."""
+    # TODO: Add support for search
+    _, key, val, created_at, updated_at = row
+    return SearchItem(
+        value=val if isinstance(val, dict) else json.loads(val),
+        key=key,
+        namespace=namespace,
+        created_at=created_at,
+        updated_at=updated_at,
+    )
+
+
 def _group_ops(ops: Iterable[Op]) -> tuple[dict[type, list[tuple[int, Op]]], int]:
    grouped_ops: dict[type, list[tuple[int, Op]]] = defaultdict(list)
    tot = 0
@@ -5,7 +5,11 @@
 ######################

 start-postgres:
-	POSTGRES_VERSION=${POSTGRES_VERSION:-16} docker compose -f tests/compose-postgres.yml up -V --force-recreate --wait
+	POSTGRES_VERSION=${POSTGRES_VERSION:-16} docker compose -f tests/compose-postgres.yml up -V --force-recreate --wait || ( \
+		echo "Failed to start PostgreSQL, printing logs..."; \
+		docker compose -f tests/compose-postgres.yml logs; \
+		exit 1 \
+	)

 stop-postgres:
 	docker compose -f tests/compose-postgres.yml down
@@ -1,10 +1,10 @@
 import threading
+from collections.abc import Iterator, Sequence
 from contextlib import contextmanager
-from typing import Any, Iterator, Optional, Sequence, Union
+from typing import Any, Optional

 from langchain_core.runnables import RunnableConfig
 from psycopg import Capabilities, Connection, Cursor, Pipeline
-from psycopg.errors import UndefinedTable
 from psycopg.rows import DictRow, dict_row
 from psycopg.types.json import Jsonb
 from psycopg_pool import ConnectionPool
@@ -17,21 +17,11 @@ from langgraph.checkpoint.base import (
    CheckpointTuple,
    get_checkpoint_id,
 )
+from langgraph.checkpoint.postgres import _internal
 from langgraph.checkpoint.postgres.base import BasePostgresSaver
 from langgraph.checkpoint.serde.base import SerializerProtocol

-Conn = Union[Connection[DictRow], ConnectionPool[Connection[DictRow]]]
-
-
-@contextmanager
-def _get_connection(conn: Conn) -> Iterator[Connection[DictRow]]:
-    if isinstance(conn, Connection):
-        yield conn
-    elif isinstance(conn, ConnectionPool):
-        with conn.connection() as conn:
-            yield conn
-    else:
-        raise TypeError(f"Invalid connection type: {type(conn)}")
+Conn = _internal.Conn  # For backward compatibility


 class PostgresSaver(BasePostgresSaver):
@@ -39,7 +29,7 @@ class PostgresSaver(BasePostgresSaver):

    def __init__(
        self,
-        conn: Conn,
+        conn: _internal.Conn,
        pipe: Optional[Pipeline] = None,
        serde: Optional[SerializerProtocol] = None,
    ) -> None:
@@ -73,9 +63,9 @@ class PostgresSaver(BasePostgresSaver):
        ) as conn:
            if pipeline:
                with conn.pipeline() as pipe:
-                    yield PostgresSaver(conn, pipe)
+                    yield cls(conn, pipe)
            else:
-                yield PostgresSaver(conn)
+                yield cls(conn)

    def setup(self) -> None:
        """Set up the checkpoint database asynchronously.
@@ -85,16 +75,15 @@ class PostgresSaver(BasePostgresSaver):
        the first time checkpointer is used.
        """
        with self._cursor() as cur:
-            try:
-                row = cur.execute(
-                    "SELECT v FROM checkpoint_migrations ORDER BY v DESC LIMIT 1"
-                ).fetchone()
-                if row is None:
-                    version = -1
-                else:
-                    version = row["v"]
-            except UndefinedTable:
+            cur.execute(self.MIGRATIONS[0])
+            results = cur.execute(
+                "SELECT v FROM checkpoint_migrations ORDER BY v DESC LIMIT 1"
+            )
+            row = results.fetchone()
+            if row is None:
                version = -1
+            else:
+                version = row["v"]
            for v, migration in zip(
                range(version + 1, len(self.MIGRATIONS)),
                self.MIGRATIONS[version + 1 :],
@@ -373,7 +362,7 @@ class PostgresSaver(BasePostgresSaver):
                Will be applied regardless of whether the PostgresSaver instance was initialized with a pipeline.
                If pipeline mode is not supported, will fall back to using transaction context manager.
        """
-        with _get_connection(self.conn) as conn:
+        with _internal.get_connection(self.conn) as conn:
            if self.pipe:
                # a connection in pipeline mode can be used concurrently
                # in multiple threads/coroutines, but only one cursor can be
@@ -388,19 +377,23 @@ class PostgresSaver(BasePostgresSaver):
                # a connection not in pipeline mode can only be used by one
                # thread/coroutine at a time, so we acquire a lock
                if self.supports_pipeline:
-                    with self.lock, conn.pipeline(), conn.cursor(
-                        binary=True, row_factory=dict_row
-                    ) as cur:
+                    with (
+                        self.lock,
+                        conn.pipeline(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
                        yield cur
                else:
                    # Use connection's transaction context manager when pipeline mode not supported
-                    with self.lock, conn.transaction(), conn.cursor(
-                        binary=True, row_factory=dict_row
-                    ) as cur:
+                    with (
+                        self.lock,
+                        conn.transaction(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
                        yield cur
            else:
                with self.lock, conn.cursor(binary=True, row_factory=dict_row) as cur:
                    yield cur


-__all__ = ["PostgresSaver", "Conn"]
+__all__ = ["PostgresSaver", "BasePostgresSaver", "Conn"]
@@ -0,0 +1,24 @@
+"""Shared async utility functions for the Postgres checkpoint & storage classes."""
+
+from collections.abc import AsyncIterator
+from contextlib import asynccontextmanager
+from typing import Union
+
+from psycopg import AsyncConnection
+from psycopg.rows import DictRow
+from psycopg_pool import AsyncConnectionPool
+
+Conn = Union[AsyncConnection[DictRow], AsyncConnectionPool[AsyncConnection[DictRow]]]
+
+
+@asynccontextmanager
+async def get_connection(
+    conn: Conn,
+) -> AsyncIterator[AsyncConnection[DictRow]]:
+    if isinstance(conn, AsyncConnection):
+        yield conn
+    elif isinstance(conn, AsyncConnectionPool):
+        async with conn.connection() as conn:
+            yield conn
+    else:
+        raise TypeError(f"Invalid connection type: {type(conn)}")
@@ -0,0 +1,22 @@
+"""Shared utility functions for the Postgres checkpoint & storage classes."""
+
+from collections.abc import Iterator
+from contextlib import contextmanager
+from typing import Union
+
+from psycopg import Connection
+from psycopg.rows import DictRow
+from psycopg_pool import ConnectionPool
+
+Conn = Union[Connection[DictRow], ConnectionPool[Connection[DictRow]]]
+
+
+@contextmanager
+def get_connection(conn: Conn) -> Iterator[Connection[DictRow]]:
+    if isinstance(conn, Connection):
+        yield conn
+    elif isinstance(conn, ConnectionPool):
+        with conn.connection() as conn:
+            yield conn
+    else:
+        raise TypeError(f"Invalid connection type: {type(conn)}")
@@ -1,10 +1,10 @@
 import asyncio
+from collections.abc import AsyncIterator, Iterator, Sequence
 from contextlib import asynccontextmanager
-from typing import Any, AsyncIterator, Iterator, Optional, Sequence, Union
+from typing import Any, Optional

 from langchain_core.runnables import RunnableConfig
 from psycopg import AsyncConnection, AsyncCursor, AsyncPipeline, Capabilities
-from psycopg.errors import UndefinedTable
 from psycopg.rows import DictRow, dict_row
 from psycopg.types.json import Jsonb
 from psycopg_pool import AsyncConnectionPool
@@ -17,23 +17,11 @@ from langgraph.checkpoint.base import (
    CheckpointTuple,
    get_checkpoint_id,
 )
+from langgraph.checkpoint.postgres import _ainternal
 from langgraph.checkpoint.postgres.base import BasePostgresSaver
 from langgraph.checkpoint.serde.base import SerializerProtocol

-Conn = Union[AsyncConnection[DictRow], AsyncConnectionPool[AsyncConnection[DictRow]]]
-
-
-@asynccontextmanager
-async def _get_connection(
-    conn: Conn,
-) -> AsyncIterator[AsyncConnection[DictRow]]:
-    if isinstance(conn, AsyncConnection):
-        yield conn
-    elif isinstance(conn, AsyncConnectionPool):
-        async with conn.connection() as conn:
-            yield conn
-    else:
-        raise TypeError(f"Invalid connection type: {type(conn)}")
+Conn = _ainternal.Conn  # For backward compatibility


 class AsyncPostgresSaver(BasePostgresSaver):
@@ -41,7 +29,7 @@ class AsyncPostgresSaver(BasePostgresSaver):

    def __init__(
        self,
-        conn: Conn,
+        conn: _ainternal.Conn,
        pipe: Optional[AsyncPipeline] = None,
        serde: Optional[SerializerProtocol] = None,
    ) -> None:
@@ -66,7 +54,7 @@ class AsyncPostgresSaver(BasePostgresSaver):
        pipeline: bool = False,
        serde: Optional[SerializerProtocol] = None,
    ) -> AsyncIterator["AsyncPostgresSaver"]:
-        """Create a new PostgresSaver instance from a connection string.
+        """Create a new AsyncPostgresSaver instance from a connection string.

        Args:
            conn_string (str): The Postgres connection info string.
@@ -80,9 +68,9 @@ class AsyncPostgresSaver(BasePostgresSaver):
        ) as conn:
            if pipeline:
                async with conn.pipeline() as pipe:
-                    yield AsyncPostgresSaver(conn=conn, pipe=pipe, serde=serde)
+                    yield cls(conn=conn, pipe=pipe, serde=serde)
            else:
-                yield AsyncPostgresSaver(conn=conn, serde=serde)
+                yield cls(conn=conn, serde=serde)

    async def setup(self) -> None:
        """Set up the checkpoint database asynchronously.
@@ -92,17 +80,15 @@ class AsyncPostgresSaver(BasePostgresSaver):
        the first time checkpointer is used.
        """
        async with self._cursor() as cur:
-            try:
-                results = await cur.execute(
-                    "SELECT v FROM checkpoint_migrations ORDER BY v DESC LIMIT 1"
-                )
-                row = await results.fetchone()
-                if row is None:
-                    version = -1
-                else:
-                    version = row["v"]
-            except UndefinedTable:
+            await cur.execute(self.MIGRATIONS[0])
+            results = await cur.execute(
+                "SELECT v FROM checkpoint_migrations ORDER BY v DESC LIMIT 1"
+            )
+            row = await results.fetchone()
+            if row is None:
                version = -1
+            else:
+                version = row["v"]
            for v, migration in zip(
                range(version + 1, len(self.MIGRATIONS)),
                self.MIGRATIONS[version + 1 :],
@@ -157,15 +143,17 @@ class AsyncPostgresSaver(BasePostgresSaver):
                        value["pending_sends"],
                    ),
                    self._load_metadata(value["metadata"]),
-                    {
-                        "configurable": {
-                            "thread_id": value["thread_id"],
-                            "checkpoint_ns": value["checkpoint_ns"],
-                            "checkpoint_id": value["parent_checkpoint_id"],
+                    (
+                        {
+                            "configurable": {
+                                "thread_id": value["thread_id"],
+                                "checkpoint_ns": value["checkpoint_ns"],
+                                "checkpoint_id": value["parent_checkpoint_id"],
+                            }
                        }
-                    }
-                    if value["parent_checkpoint_id"]
-                    else None,
+                        if value["parent_checkpoint_id"]
+                        else None
+                    ),
                    await asyncio.to_thread(self._load_writes, value["pending_writes"]),
                )

@@ -216,15 +204,17 @@ class AsyncPostgresSaver(BasePostgresSaver):
                        value["pending_sends"],
                    ),
                    self._load_metadata(value["metadata"]),
-                    {
-                        "configurable": {
-                            "thread_id": thread_id,
-                            "checkpoint_ns": checkpoint_ns,
-                            "checkpoint_id": value["parent_checkpoint_id"],
+                    (
+                        {
+                            "configurable": {
+                                "thread_id": thread_id,
+                                "checkpoint_ns": checkpoint_ns,
+                                "checkpoint_id": value["parent_checkpoint_id"],
+                            }
                        }
-                    }
-                    if value["parent_checkpoint_id"]
-                    else None,
+                        if value["parent_checkpoint_id"]
+                        else None
+                    ),
                    await asyncio.to_thread(self._load_writes, value["pending_writes"]),
                )

@@ -331,7 +321,7 @@ class AsyncPostgresSaver(BasePostgresSaver):
                Will be applied regardless of whether the AsyncPostgresSaver instance was initialized with a pipeline.
                If pipeline mode is not supported, will fall back to using transaction context manager.
        """
-        async with _get_connection(self.conn) as conn:
+        async with _ainternal.get_connection(self.conn) as conn:
            if self.pipe:
                # a connection in pipeline mode can be used concurrently
                # in multiple threads/coroutines, but only one cursor can be
@@ -346,20 +336,25 @@ class AsyncPostgresSaver(BasePostgresSaver):
                # a connection not in pipeline mode can only be used by one
                # thread/coroutine at a time, so we acquire a lock
                if self.supports_pipeline:
-                    async with self.lock, conn.pipeline(), conn.cursor(
-                        binary=True, row_factory=dict_row
-                    ) as cur:
+                    async with (
+                        self.lock,
+                        conn.pipeline(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
                        yield cur
                else:
                    # Use connection's transaction context manager when pipeline mode not supported
-                    async with self.lock, conn.transaction(), conn.cursor(
-                        binary=True, row_factory=dict_row
-                    ) as cur:
+                    async with (
+                        self.lock,
+                        conn.transaction(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
                        yield cur
            else:
-                async with self.lock, conn.cursor(
-                    binary=True, row_factory=dict_row
-                ) as cur:
+                async with (
+                    self.lock,
+                    conn.cursor(binary=True, row_factory=dict_row) as cur,
+                ):
                    yield cur

    def list(
@@ -388,7 +383,7 @@ class AsyncPostgresSaver(BasePostgresSaver):
        while True:
            try:
                yield asyncio.run_coroutine_threadsafe(
-                    anext(aiter_),
+                    anext(aiter_),  # noqa: F821
                    self.loop,
                ).result()
            except StopAsyncIteration:
@@ -467,3 +462,6 @@ class AsyncPostgresSaver(BasePostgresSaver):
        return asyncio.run_coroutine_threadsafe(
            self.aput_writes(config, writes, task_id), self.loop
        ).result()
+
+
+__all__ = ["AsyncPostgresSaver", "Conn"]
@@ -1,5 +1,6 @@
 import random
-from typing import Any, List, Optional, Sequence, Tuple, cast
+from collections.abc import Sequence
+from typing import Any, Optional, cast

 from langchain_core.runnables import RunnableConfig
 from psycopg.types.json import Jsonb
@@ -56,6 +57,17 @@ MIGRATIONS = [
    PRIMARY KEY (thread_id, checkpoint_ns, checkpoint_id, task_id, idx)
 );""",
    "ALTER TABLE checkpoint_blobs ALTER COLUMN blob DROP not null;",
+    """
+    """,
+    """
+    CREATE INDEX CONCURRENTLY IF NOT EXISTS checkpoints_thread_id_idx ON checkpoints(thread_id);
+    """,
+    """
+    CREATE INDEX CONCURRENTLY IF NOT EXISTS checkpoint_blobs_thread_id_idx ON checkpoint_blobs(thread_id);
+    """,
+    """
+    CREATE INDEX CONCURRENTLY IF NOT EXISTS checkpoint_writes_thread_id_idx ON checkpoint_writes(thread_id);
+    """,
 ]

 SELECT_SQL = f"""
@@ -249,7 +261,7 @@ class BasePostgresSaver(BaseCheckpointSaver[str]):
        config: Optional[RunnableConfig],
        filter: MetadataInput,
        before: Optional[RunnableConfig] = None,
-    ) -> Tuple[str, List[Any]]:
+    ) -> tuple[str, list[Any]]:
        """Return WHERE clause predicates for alist() given config, filter, before.

        This method returns a tuple of a string and a tuple of values. The string
@@ -1,110 +1,303 @@
 import asyncio
 import logging
+from collections.abc import AsyncIterator, Iterable, Sequence
 from contextlib import asynccontextmanager
-from typing import (
-    Any,
-    AsyncIterator,
-    Callable,
-    Iterable,
-    Optional,
-    Sequence,
-    Union,
-    cast,
-)
+from typing import Any, Callable, Optional, Union, cast

 import orjson
-from psycopg import AsyncConnection, AsyncCursor
-from psycopg.errors import UndefinedTable
-from psycopg.rows import dict_row
+from psycopg import AsyncConnection, AsyncCursor, AsyncPipeline, Capabilities
+from psycopg.rows import DictRow, dict_row
+from psycopg_pool import AsyncConnectionPool

-from langgraph.store.base import GetOp, ListNamespacesOp, Op, PutOp, Result, SearchOp
+from langgraph.checkpoint.postgres import _ainternal
+from langgraph.store.base import (
+    GetOp,
+    ListNamespacesOp,
+    Op,
+    PutOp,
+    Result,
+    SearchOp,
+)
 from langgraph.store.base.batch import AsyncBatchedBaseStore
 from langgraph.store.postgres.base import (
+    _PLACEHOLDER,
    BasePostgresStore,
+    PoolConfig,
+    PostgresIndexConfig,
    Row,
    _decode_ns_bytes,
+    _ensure_index_config,
    _group_ops,
    _row_to_item,
+    _row_to_search_item,
 )

 logger = logging.getLogger(__name__)


-class AsyncPostgresStore(AsyncBatchedBaseStore, BasePostgresStore[AsyncConnection]):
-    __slots__ = ("_deserializer",)
+class AsyncPostgresStore(AsyncBatchedBaseStore, BasePostgresStore[_ainternal.Conn]):
+    """Asynchronous Postgres-backed store with optional vector search using pgvector.
+
+    !!! example "Examples"
+        Basic setup and key-value storage:
+        ```python
+        from langgraph.store.postgres import AsyncPostgresStore
+
+        async with AsyncPostgresStore.from_conn_string(
+            "postgresql://user:pass@localhost:5432/dbname"
+        ) as store:
+            await store.setup()
+
+            # Store and retrieve data
+            await store.aput(("users", "123"), "prefs", {"theme": "dark"})
+            item = await store.aget(("users", "123"), "prefs")
+        ```
+
+        Vector search using LangChain embeddings:
+        ```python
+        from langchain.embeddings import init_embeddings
+        from langgraph.store.postgres import AsyncPostgresStore
+
+        async with AsyncPostgresStore.from_conn_string(
+            "postgresql://user:pass@localhost:5432/dbname",
+            index={
+                "dims": 1536,
+                "embed": init_embeddings("openai:text-embedding-3-small"),
+                "fields": ["text"]  # specify which fields to embed. Default is the whole serialized value
+            }
+        ) as store:
+            await store.setup()  # Do this once to run migrations
+
+            # Store documents
+            await store.aput(("docs",), "doc1", {"text": "Python tutorial"})
+            await store.aput(("docs",), "doc2", {"text": "TypeScript guide"})
+            # Don't index the following
+            await store.aput(("docs",), "doc3", {"text": "Other guide"}, index=False)
+
+            # Search by similarity
+            results = await store.asearch(("docs",), query="python programming")
+        ```
+
+        Using connection pooling for better performance:
+        ```python
+        from langgraph.store.postgres import AsyncPostgresStore, PoolConfig
+
+        async with AsyncPostgresStore.from_conn_string(
+            "postgresql://user:pass@localhost:5432/dbname",
+            pool_config=PoolConfig(
+                min_size=5,
+                max_size=20
+            )
+        ) as store:
+            await store.setup()
+            # Use store with connection pooling...
+        ```
+
+    Warning:
+        Make sure to:
+        1. Call `setup()` before first use to create necessary tables and indexes
+        2. Have the pgvector extension available to use vector search
+        3. Use Python 3.10+ for async functionality
+
+    Note:
+        Semantic search is disabled by default. You can enable it by providing an `index` configuration
+        when creating the store. Without this configuration, all `index` arguments passed to
+        `put` or `aput`will have no effect.
+    """
+
+    __slots__ = (
+        "_deserializer",
+        "pipe",
+        "lock",
+        "supports_pipeline",
+        "index_config",
+        "embeddings",
+    )

    def __init__(
        self,
-        conn: AsyncConnection[Any],
+        conn: _ainternal.Conn,
        *,
+        pipe: Optional[AsyncPipeline] = None,
        deserializer: Optional[
            Callable[[Union[bytes, orjson.Fragment]], dict[str, Any]]
        ] = None,
+        index: Optional[PostgresIndexConfig] = None,
    ) -> None:
+        if isinstance(conn, AsyncConnectionPool) and pipe is not None:
+            raise ValueError(
+                "Pipeline should be used only with a single AsyncConnection, not AsyncConnectionPool."
+            )
        super().__init__()
        self._deserializer = deserializer
        self.conn = conn
+        self.pipe = pipe
+        self.lock = asyncio.Lock()
        self.loop = asyncio.get_running_loop()
+        self.supports_pipeline = Capabilities().has_pipeline()
+        self.index_config = index
+        if self.index_config:
+            self.embeddings, self.index_config = _ensure_index_config(self.index_config)
+
+        else:
+            self.embeddings = None

    async def abatch(self, ops: Iterable[Op]) -> list[Result]:
        grouped_ops, num_ops = _group_ops(ops)
        results: list[Result] = [None] * num_ops

-        async with self.conn.pipeline():
-            tasks = []
-
-            if GetOp in grouped_ops:
-                tasks.append(
-                    self._batch_get_ops(
-                        cast(Sequence[tuple[int, GetOp]], grouped_ops[GetOp]), results
-                    )
-                )
-
-            if PutOp in grouped_ops:
-                tasks.append(
-                    self._batch_put_ops(
-                        cast(Sequence[tuple[int, PutOp]], grouped_ops[PutOp])
-                    )
-                )
-
-            if SearchOp in grouped_ops:
-                tasks.append(
-                    self._batch_search_ops(
-                        cast(Sequence[tuple[int, SearchOp]], grouped_ops[SearchOp]),
-                        results,
-                    )
-                )
-
-            if ListNamespacesOp in grouped_ops:
-                tasks.append(
-                    self._batch_list_namespaces_ops(
-                        cast(
-                            Sequence[tuple[int, ListNamespacesOp]],
-                            grouped_ops[ListNamespacesOp],
-                        ),
-                        results,
-                    )
-                )
-
-            await asyncio.gather(*tasks)
+        async with _ainternal.get_connection(self.conn) as conn:
+            if self.pipe:
+                async with self.pipe:
+                    await self._execute_batch(grouped_ops, results, conn)
+            else:
+                await self._execute_batch(grouped_ops, results, conn)

        return results

-    def batch(self, ops: Iterable[Op]) -> list[Result]:
-        return asyncio.run_coroutine_threadsafe(self.abatch(ops), self.loop).result()
+    @classmethod
+    @asynccontextmanager
+    async def from_conn_string(
+        cls,
+        conn_string: str,
+        *,
+        pipeline: bool = False,
+        pool_config: Optional[PoolConfig] = None,
+        index: Optional[PostgresIndexConfig] = None,
+    ) -> AsyncIterator["AsyncPostgresStore"]:
+        """Create a new AsyncPostgresStore instance from a connection string.
+
+        Args:
+            conn_string (str): The Postgres connection info string.
+            pipeline (bool): Whether to use AsyncPipeline (only for single connections)
+            pool_config (Optional[PoolConfig]): Configuration for the connection pool.
+                If provided, will create a connection pool and use it instead of a single connection.
+                This overrides the `pipeline` argument.
+            index (Optional[PostgresIndexConfig]): The embedding config.
+
+        Returns:
+            AsyncPostgresStore: A new AsyncPostgresStore instance.
+        """
+        if pool_config is not None:
+            pc = pool_config.copy()
+            async with cast(
+                AsyncConnectionPool[AsyncConnection[DictRow]],
+                AsyncConnectionPool(
+                    conn_string,
+                    min_size=pc.pop("min_size", 1),
+                    max_size=pc.pop("max_size", None),
+                    kwargs={
+                        "autocommit": True,
+                        "prepare_threshold": 0,
+                        "row_factory": dict_row,
+                        **(pc.pop("kwargs", None) or {}),
+                    },
+                    **cast(dict, pc),
+                ),
+            ) as pool:
+                yield cls(conn=pool, index=index)
+        else:
+            async with await AsyncConnection.connect(
+                conn_string, autocommit=True, prepare_threshold=0, row_factory=dict_row
+            ) as conn:
+                if pipeline:
+                    async with conn.pipeline() as pipe:
+                        yield cls(conn=conn, pipe=pipe, index=index)
+                else:
+                    yield cls(conn=conn, index=index)
+
+    async def setup(self) -> None:
+        """Set up the store database asynchronously.
+
+        This method creates the necessary tables in the Postgres database if they don't
+        already exist and runs database migrations. It MUST be called directly by the user
+        the first time the store is used.
+        """
+
+        async def _get_version(cur: AsyncCursor[DictRow], table: str) -> int:
+            await cur.execute(
+                f"""
+                CREATE TABLE IF NOT EXISTS {table} (
+                    v INTEGER PRIMARY KEY
+                )
+            """
+            )
+            await cur.execute(f"SELECT v FROM {table} ORDER BY v DESC LIMIT 1")
+            row = cast(dict, await cur.fetchone())
+            if row is None:
+                version = -1
+            else:
+                version = row["v"]
+            return version
+
+        async with self._cursor() as cur:
+            version = await _get_version(cur, table="store_migrations")
+            for v, sql in enumerate(self.MIGRATIONS[version + 1 :], start=version + 1):
+                await cur.execute(sql)
+                await cur.execute("INSERT INTO store_migrations (v) VALUES (%s)", (v,))
+
+            if self.index_config:
+                version = await _get_version(cur, table="vector_migrations")
+                for v, migration in enumerate(
+                    self.VECTOR_MIGRATIONS[version + 1 :], start=version + 1
+                ):
+                    sql = migration.sql
+                    if migration.params:
+                        params = {
+                            k: v(self) if v is not None and callable(v) else v
+                            for k, v in migration.params.items()
+                        }
+                        sql = sql % params
+                    await cur.execute(sql)
+                    await cur.execute(
+                        "INSERT INTO vector_migrations (v) VALUES (%s)", (v,)
+                    )
+
+    async def _execute_batch(
+        self,
+        grouped_ops: dict,
+        results: list[Result],
+        conn: AsyncConnection[DictRow],
+    ) -> None:
+        async with self._cursor(pipeline=True) as cur:
+            if GetOp in grouped_ops:
+                await self._batch_get_ops(
+                    cast(Sequence[tuple[int, GetOp]], grouped_ops[GetOp]),
+                    results,
+                    cur,
+                )
+
+            if SearchOp in grouped_ops:
+                await self._batch_search_ops(
+                    cast(Sequence[tuple[int, SearchOp]], grouped_ops[SearchOp]),
+                    results,
+                    cur,
+                )
+
+            if ListNamespacesOp in grouped_ops:
+                await self._batch_list_namespaces_ops(
+                    cast(
+                        Sequence[tuple[int, ListNamespacesOp]],
+                        grouped_ops[ListNamespacesOp],
+                    ),
+                    results,
+                    cur,
+                )
+
+            if PutOp in grouped_ops:
+                await self._batch_put_ops(
+                    cast(Sequence[tuple[int, PutOp]], grouped_ops[PutOp]),
+                    cur,
+                )

    async def _batch_get_ops(
        self,
        get_ops: Sequence[tuple[int, GetOp]],
        results: list[Result],
+        cur: AsyncCursor[DictRow],
    ) -> None:
-        cursors = []
        for query, params, namespace, items in self._get_batch_GET_ops_queries(get_ops):
-            cur = self.conn.cursor(binary=True)
            await cur.execute(query, params)
-            cursors.append((cur, namespace, items))
-
-        for cur, namespace, items in cursors:
            rows = cast(list[Row], await cur.fetchall())
            key_to_row = {row["key"]: row for row in rows}
            for idx, key in items:
@@ -119,29 +312,59 @@ class AsyncPostgresStore(AsyncBatchedBaseStore, BasePostgresStore[AsyncConnectio
    async def _batch_put_ops(
        self,
        put_ops: Sequence[tuple[int, PutOp]],
+        cur: AsyncCursor[DictRow],
    ) -> None:
-        queries = self._get_batch_PUT_queries(put_ops)
+        queries, embedding_request = self._prepare_batch_PUT_queries(put_ops)
+        if embedding_request:
+            if self.embeddings is None:
+                # Should not get here since the embedding config is required
+                # to return an embedding_request above
+                raise ValueError(
+                    "Embedding configuration is required for vector operations "
+                    f"(for semantic search). "
+                    f"Please provide an EmbeddingConfig when initializing the {self.__class__.__name__}."
+                )
+            query, txt_params = embedding_request
+            vectors = await self.embeddings.aembed_documents(
+                [param[-1] for param in txt_params]
+            )
+            queries.append(
+                (
+                    query,
+                    [
+                        p
+                        for (ns, k, pathname, _), vector in zip(txt_params, vectors)
+                        for p in (ns, k, pathname, vector)
+                    ],
+                )
+            )
+
        for query, params in queries:
-            cur = self.conn.cursor(binary=True)
            await cur.execute(query, params)

    async def _batch_search_ops(
        self,
        search_ops: Sequence[tuple[int, SearchOp]],
        results: list[Result],
+        cur: AsyncCursor[DictRow],
    ) -> None:
-        queries = self._get_batch_search_queries(search_ops)
-        cursors: list[tuple[AsyncCursor[Any], int]] = []
+        queries, embedding_requests = self._prepare_batch_search_queries(search_ops)

-        for (query, params), (idx, _) in zip(queries, search_ops):
-            cur = self.conn.cursor(binary=True)
+        if embedding_requests and self.embeddings:
+            vectors = await self.embeddings.aembed_documents(
+                [query for _, query in embedding_requests]
+            )
+            for (idx, _), vector in zip(embedding_requests, vectors):
+                _paramslist = queries[idx][1]
+                for i in range(len(_paramslist)):
+                    if _paramslist[i] is _PLACEHOLDER:
+                        _paramslist[i] = vector
+
+        for (idx, _), (query, params) in zip(search_ops, queries):
            await cur.execute(query, params)
-            cursors.append((cur, idx))
-
-        for cur, idx in cursors:
            rows = cast(list[Row], await cur.fetchall())
            items = [
-                _row_to_item(
+                _row_to_search_item(
                    _decode_ns_bytes(row["prefix"]), row, loader=self._deserializer
                )
                for row in rows
@@ -152,67 +375,57 @@ class AsyncPostgresStore(AsyncBatchedBaseStore, BasePostgresStore[AsyncConnectio
        self,
        list_ops: Sequence[tuple[int, ListNamespacesOp]],
        results: list[Result],
+        cur: AsyncCursor[DictRow],
    ) -> None:
        queries = self._get_batch_list_namespaces_queries(list_ops)
-        cursors: list[tuple[AsyncCursor[Any], int]] = []
        for (query, params), (idx, _) in zip(queries, list_ops):
-            cur = self.conn.cursor(binary=True)
            await cur.execute(query, params)
-            cursors.append((cur, idx))
-
-        for cur, idx in cursors:
            rows = cast(list[dict], await cur.fetchall())
            namespaces = [_decode_ns_bytes(row["truncated_prefix"]) for row in rows]
            results[idx] = namespaces

-    @classmethod
    @asynccontextmanager
-    async def from_conn_string(
-        cls,
-        conn_string: str,
-    ) -> AsyncIterator["AsyncPostgresStore"]:
-        """Create a new AsyncPostgresStore instance from a connection string.
+    async def _cursor(
+        self, *, pipeline: bool = False
+    ) -> AsyncIterator[AsyncCursor[DictRow]]:
+        """Create a database cursor as a context manager.

        Args:
-            conn_string (str): The Postgres connection info string.
-
-        Returns:
-            AsyncPostgresStore: A new AsyncPostgresStore instance.
+            pipeline: whether to use pipeline for the DB operations inside the context manager.
+                Will be applied regardless of whether the PostgresStore instance was initialized with a pipeline.
+                If pipeline mode is not supported, will fall back to using transaction context manager.
        """
-        async with await AsyncConnection.connect(
-            conn_string, autocommit=True, prepare_threshold=0, row_factory=dict_row
-        ) as conn:
-            yield cls(conn=conn)
-
-    async def setup(self) -> None:
-        """Set up the store database asynchronously.
-
-        This method creates the necessary tables in the Postgres database if they don't
-        already exist and runs database migrations. It MUST be called directly by the user
-        the first time the store is used.
-        """
-        async with self.conn.cursor() as cur:
-            try:
-                await cur.execute(
-                    "SELECT v FROM store_migrations ORDER BY v DESC LIMIT 1"
-                )
-                row = cast(dict, await cur.fetchone())
-                if row is None:
-                    version = -1
+        async with _ainternal.get_connection(self.conn) as conn:
+            if self.pipe:
+                # a connection in pipeline mode can be used concurrently
+                # in multiple threads/coroutines, but only one cursor can be
+                # used at a time
+                try:
+                    async with conn.cursor(binary=True, row_factory=dict_row) as cur:
+                        yield cur
+                finally:
+                    if pipeline:
+                        await self.pipe.sync()
+            elif pipeline:
+                # a connection not in pipeline mode can only be used by one
+                # thread/coroutine at a time, so we acquire a lock
+                if self.supports_pipeline:
+                    async with (
+                        self.lock,
+                        conn.pipeline(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
+                        yield cur
                else:
-                    version = row["v"]
-            except UndefinedTable:
-                version = -1
-                # Create store_migrations table if it doesn't exist
-                await cur.execute(
-                    """
-                    CREATE TABLE IF NOT EXISTS store_migrations (
-                        v INTEGER PRIMARY KEY
-                    )
-                    """
-                )
-            for v, migration in enumerate(
-                self.MIGRATIONS[version + 1 :], start=version + 1
-            ):
-                await cur.execute(migration)
-                await cur.execute("INSERT INTO store_migrations (v) VALUES (%s)", (v,))
+                    async with (
+                        self.lock,
+                        conn.transaction(),
+                        conn.cursor(binary=True, row_factory=dict_row) as cur,
+                    ):
+                        yield cur
+            else:
+                async with (
+                    self.lock,
+                    conn.cursor(binary=True) as cur,
+                ):
+                    yield cur
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "langgraph-checkpoint-postgres"
-version = "2.0.3"
+version = "2.0.8"
 description = "Library with a Postgres implementation of LangGraph checkpoint saver."
 authors = []
 license = "MIT"
@@ -10,10 +10,10 @@ packages = [{ include = "langgraph" }]

 [tool.poetry.dependencies]
 python = "^3.9.0,<4.0"
-langgraph-checkpoint = "^2.0.2"
+langgraph-checkpoint = "^2.0.7"
 orjson = ">=3.10.1"
-psycopg = "^3.0.0"
-psycopg-pool = "^3.0.0"
+psycopg = "^3.2.0"
+psycopg-pool = "^3.2.0"

 [tool.poetry.group.dev.dependencies]
 ruff = "^0.6.2"
@@ -1,12 +1,13 @@
 services:
  postgres-test:
-    image: postgres:${POSTGRES_VERSION:-16}
+    image: pgvector/pgvector:pg${POSTGRES_VERSION:-16}
    ports:
      - "5441:5432"
    environment:
      POSTGRES_DB: postgres
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
+    command: ["postgres", "-c", "shared_preload_libraries=vector"]
    healthcheck:
      test: pg_isready -U postgres
      start_period: 10s
@@ -1,10 +1,13 @@
-from typing import AsyncIterator
+from collections.abc import AsyncIterator

 import pytest
 from psycopg import AsyncConnection
 from psycopg.errors import UndefinedTable
 from psycopg.rows import DictRow, dict_row

+from tests.embed_test_utils import CharacterEmbeddings
+
+DEFAULT_POSTGRES_URI = "postgres://postgres:postgres@localhost:5441/"
 DEFAULT_URI = "postgres://postgres:postgres@localhost:5441/postgres?sslmode=disable"


@@ -24,6 +27,18 @@ async def clear_test_db(conn: AsyncConnection[DictRow]) -> None:
        await conn.execute("DELETE FROM checkpoint_blobs")
        await conn.execute("DELETE FROM checkpoint_writes")
        await conn.execute("DELETE FROM checkpoint_migrations")
-        await conn.execute("DELETE FROM store_migrations")
    except UndefinedTable:
        pass
+    try:
+        await conn.execute("DELETE FROM store_migrations")
+        await conn.execute("DELETE FROM store")
+    except UndefinedTable:
+        pass
+
+
+@pytest.fixture
+def fake_embeddings() -> CharacterEmbeddings:
+    return CharacterEmbeddings(dims=500)
+
+
+VECTOR_TYPES = ["vector", "halfvec"]
@@ -0,0 +1,55 @@
+"""Embedding utilities for testing."""
+
+import math
+import random
+from collections import Counter, defaultdict
+from typing import Any
+
+from langchain_core.embeddings import Embeddings
+
+
+class CharacterEmbeddings(Embeddings):
+    """Simple character-frequency based embeddings using random projections."""
+
+    def __init__(self, dims: int = 50, seed: int = 42):
+        """Initialize with embedding dimensions and random seed."""
+        self._rng = random.Random(seed)
+        self.dims = dims
+        # Create projection vector for each character lazily
+        self._char_projections: defaultdict[str, list[float]] = defaultdict(
+            lambda: [
+                self._rng.gauss(0, 1 / math.sqrt(self.dims)) for _ in range(self.dims)
+            ]
+        )
+
+    def _embed_one(self, text: str) -> list[float]:
+        """Embed a single text."""
+        counts = Counter(text)
+        total = sum(counts.values())
+
+        if total == 0:
+            return [0.0] * self.dims
+
+        embedding = [0.0] * self.dims
+        for char, count in counts.items():
+            weight = count / total
+            char_proj = self._char_projections[char]
+            for i, proj in enumerate(char_proj):
+                embedding[i] += weight * proj
+
+        norm = math.sqrt(sum(x * x for x in embedding))
+        if norm > 0:
+            embedding = [x / norm for x in embedding]
+
+        return embedding
+
+    def embed_documents(self, texts: list[str]) -> list[list[float]]:
+        """Embed a list of documents."""
+        return [self._embed_one(text) for text in texts]
+
+    def embed_query(self, text: str) -> list[float]:
+        """Embed a query string."""
+        return self._embed_one(text)
+
+    def __eq__(self, other: Any) -> bool:
+        return isinstance(other, CharacterEmbeddings) and self.dims == other.dims
@@ -1,8 +1,14 @@
+# type: ignore
+
+from contextlib import asynccontextmanager
 from typing import Any
+from uuid import uuid4

 import pytest
-from conftest import DEFAULT_URI  # type: ignore
 from langchain_core.runnables import RunnableConfig
+from psycopg import AsyncConnection
+from psycopg.rows import dict_row
+from psycopg_pool import AsyncConnectionPool

 from langgraph.checkpoint.base import (
    Checkpoint,
@@ -11,103 +17,211 @@ from langgraph.checkpoint.base import (
    empty_checkpoint,
 )
 from langgraph.checkpoint.postgres.aio import AsyncPostgresSaver
+from tests.conftest import DEFAULT_POSTGRES_URI


-class TestAsyncPostgresSaver:
-    @pytest.fixture(autouse=True)
-    async def setup(self) -> None:
-        # objects for test setup
-        self.config_1: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-1",
-                # for backwards compatibility testing
-                "thread_ts": "1",
-                "checkpoint_ns": "",
-            }
-        }
-        self.config_2: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-2",
-                "checkpoint_id": "2",
-                "checkpoint_ns": "",
-            }
-        }
-        self.config_3: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-2",
-                "checkpoint_id": "2-inner",
-                "checkpoint_ns": "inner",
-            }
-        }
+@asynccontextmanager
+async def _pool_saver():
+    """Fixture for pool mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    async with await AsyncConnection.connect(
+        DEFAULT_POSTGRES_URI, autocommit=True
+    ) as conn:
+        await conn.execute(f"CREATE DATABASE {database}")
+    try:
+        # yield checkpointer
+        async with AsyncConnectionPool(
+            DEFAULT_POSTGRES_URI + database,
+            max_size=10,
+            kwargs={"autocommit": True, "row_factory": dict_row},
+        ) as pool:
+            checkpointer = AsyncPostgresSaver(pool)
+            await checkpointer.setup()
+            yield checkpointer
+    finally:
+        # drop unique db
+        async with await AsyncConnection.connect(
+            DEFAULT_POSTGRES_URI, autocommit=True
+        ) as conn:
+            await conn.execute(f"DROP DATABASE {database}")

-        self.chkpnt_1: Checkpoint = empty_checkpoint()
-        self.chkpnt_2: Checkpoint = create_checkpoint(self.chkpnt_1, {}, 1)
-        self.chkpnt_3: Checkpoint = empty_checkpoint()

-        self.metadata_1: CheckpointMetadata = {
-            "source": "input",
-            "step": 2,
-            "writes": {},
-            "score": 1,
+@asynccontextmanager
+async def _pipe_saver():
+    """Fixture for pipeline mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    async with await AsyncConnection.connect(
+        DEFAULT_POSTGRES_URI, autocommit=True
+    ) as conn:
+        await conn.execute(f"CREATE DATABASE {database}")
+    try:
+        async with await AsyncConnection.connect(
+            DEFAULT_POSTGRES_URI + database,
+            autocommit=True,
+            prepare_threshold=0,
+            row_factory=dict_row,
+        ) as conn:
+            checkpointer = AsyncPostgresSaver(conn)
+            await checkpointer.setup()
+            async with conn.pipeline() as pipe:
+                checkpointer = AsyncPostgresSaver(conn, pipe=pipe)
+                yield checkpointer
+    finally:
+        # drop unique db
+        async with await AsyncConnection.connect(
+            DEFAULT_POSTGRES_URI, autocommit=True
+        ) as conn:
+            await conn.execute(f"DROP DATABASE {database}")
+
+
+@asynccontextmanager
+async def _base_saver():
+    """Fixture for regular connection mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    async with await AsyncConnection.connect(
+        DEFAULT_POSTGRES_URI, autocommit=True
+    ) as conn:
+        await conn.execute(f"CREATE DATABASE {database}")
+    try:
+        async with await AsyncConnection.connect(
+            DEFAULT_POSTGRES_URI + database,
+            autocommit=True,
+            prepare_threshold=0,
+            row_factory=dict_row,
+        ) as conn:
+            checkpointer = AsyncPostgresSaver(conn)
+            await checkpointer.setup()
+            yield checkpointer
+    finally:
+        # drop unique db
+        async with await AsyncConnection.connect(
+            DEFAULT_POSTGRES_URI, autocommit=True
+        ) as conn:
+            await conn.execute(f"DROP DATABASE {database}")
+
+
+@asynccontextmanager
+async def _saver(name: str):
+    if name == "base":
+        async with _base_saver() as saver:
+            yield saver
+    elif name == "pool":
+        async with _pool_saver() as saver:
+            yield saver
+    elif name == "pipe":
+        async with _pipe_saver() as saver:
+            yield saver
+
+
+@pytest.fixture
+def test_data():
+    """Fixture providing test data for checkpoint tests."""
+    config_1: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-1",
+            # for backwards compatibility testing
+            "thread_ts": "1",
+            "checkpoint_ns": "",
        }
-        self.metadata_2: CheckpointMetadata = {
-            "source": "loop",
+    }
+    config_2: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-2",
+            "checkpoint_id": "2",
+            "checkpoint_ns": "",
+        }
+    }
+    config_3: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-2",
+            "checkpoint_id": "2-inner",
+            "checkpoint_ns": "inner",
+        }
+    }
+
+    chkpnt_1: Checkpoint = empty_checkpoint()
+    chkpnt_2: Checkpoint = create_checkpoint(chkpnt_1, {}, 1)
+    chkpnt_3: Checkpoint = empty_checkpoint()
+
+    metadata_1: CheckpointMetadata = {
+        "source": "input",
+        "step": 2,
+        "writes": {},
+        "score": 1,
+    }
+    metadata_2: CheckpointMetadata = {
+        "source": "loop",
+        "step": 1,
+        "writes": {"foo": "bar"},
+        "score": None,
+    }
+    metadata_3: CheckpointMetadata = {}
+
+    return {
+        "configs": [config_1, config_2, config_3],
+        "checkpoints": [chkpnt_1, chkpnt_2, chkpnt_3],
+        "metadata": [metadata_1, metadata_2, metadata_3],
+    }
+
+
+@pytest.mark.parametrize("saver_name", ["base", "pool", "pipe"])
+async def test_asearch(request, saver_name: str, test_data) -> None:
+    async with _saver(saver_name) as saver:
+        configs = test_data["configs"]
+        checkpoints = test_data["checkpoints"]
+        metadata = test_data["metadata"]
+
+        await saver.aput(configs[0], checkpoints[0], metadata[0], {})
+        await saver.aput(configs[1], checkpoints[1], metadata[1], {})
+        await saver.aput(configs[2], checkpoints[2], metadata[2], {})
+
+        # call method / assertions
+        query_1 = {"source": "input"}  # search by 1 key
+        query_2 = {
            "step": 1,
            "writes": {"foo": "bar"},
-            "score": None,
-        }
-        self.metadata_3: CheckpointMetadata = {}
-        async with AsyncPostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            await saver.setup()
+        }  # search by multiple keys
+        query_3: dict[str, Any] = {}  # search by no keys, return all checkpoints
+        query_4 = {"source": "update", "step": 1}  # no match

-    async def test_asearch(self) -> None:
-        async with AsyncPostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            await saver.aput(self.config_1, self.chkpnt_1, self.metadata_1, {})
-            await saver.aput(self.config_2, self.chkpnt_2, self.metadata_2, {})
-            await saver.aput(self.config_3, self.chkpnt_3, self.metadata_3, {})
+        search_results_1 = [c async for c in saver.alist(None, filter=query_1)]
+        assert len(search_results_1) == 1
+        assert search_results_1[0].metadata == metadata[0]

-            # call method / assertions
-            query_1 = {"source": "input"}  # search by 1 key
-            query_2 = {
-                "step": 1,
-                "writes": {"foo": "bar"},
-            }  # search by multiple keys
-            query_3: dict[str, Any] = {}  # search by no keys, return all checkpoints
-            query_4 = {"source": "update", "step": 1}  # no match
+        search_results_2 = [c async for c in saver.alist(None, filter=query_2)]
+        assert len(search_results_2) == 1
+        assert search_results_2[0].metadata == metadata[1]

-            search_results_1 = [c async for c in saver.alist(None, filter=query_1)]
-            assert len(search_results_1) == 1
-            assert search_results_1[0].metadata == self.metadata_1
+        search_results_3 = [c async for c in saver.alist(None, filter=query_3)]
+        assert len(search_results_3) == 3

-            search_results_2 = [c async for c in saver.alist(None, filter=query_2)]
-            assert len(search_results_2) == 1
-            assert search_results_2[0].metadata == self.metadata_2
+        search_results_4 = [c async for c in saver.alist(None, filter=query_4)]
+        assert len(search_results_4) == 0

-            search_results_3 = [c async for c in saver.alist(None, filter=query_3)]
-            assert len(search_results_3) == 3
+        # search by config (defaults to checkpoints across all namespaces)
+        search_results_5 = [
+            c async for c in saver.alist({"configurable": {"thread_id": "thread-2"}})
+        ]
+        assert len(search_results_5) == 2
+        assert {
+            search_results_5[0].config["configurable"]["checkpoint_ns"],
+            search_results_5[1].config["configurable"]["checkpoint_ns"],
+        } == {"", "inner"}

-            search_results_4 = [c async for c in saver.alist(None, filter=query_4)]
-            assert len(search_results_4) == 0

-            # search by config (defaults to checkpoints across all namespaces)
-            search_results_5 = [
-                c
-                async for c in saver.alist({"configurable": {"thread_id": "thread-2"}})
-            ]
-            assert len(search_results_5) == 2
-            assert {
-                search_results_5[0].config["configurable"]["checkpoint_ns"],
-                search_results_5[1].config["configurable"]["checkpoint_ns"],
-            } == {"", "inner"}
-
-            # TODO: test before and limit params
-
-    async def test_null_chars(self) -> None:
-        async with AsyncPostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            config = await saver.aput(
-                self.config_1, self.chkpnt_1, {"my_key": "\x00abc"}, {}
-            )
-            assert (await saver.aget_tuple(config)).metadata["my_key"] == "abc"  # type: ignore
-            assert [c async for c in saver.alist(None, filter={"my_key": "abc"})][
-                0
-            ].metadata["my_key"] == "abc"
+@pytest.mark.parametrize("saver_name", ["base", "pool", "pipe"])
+async def test_null_chars(request, saver_name: str, test_data) -> None:
+    async with _saver(saver_name) as saver:
+        config = await saver.aput(
+            test_data["configs"][0],
+            test_data["checkpoints"][0],
+            {"my_key": "\x00abc"},
+            {},
+        )
+        assert (await saver.aget_tuple(config)).metadata["my_key"] == "abc"  # type: ignore
+        assert [c async for c in saver.alist(None, filter={"my_key": "abc"})][
+            0
+        ].metadata["my_key"] == "abc"
@@ -1,114 +1,214 @@
 # type: ignore
+import asyncio
+import itertools
+import sys
 import uuid
-from datetime import datetime
-from typing import Any
-from unittest.mock import AsyncMock, MagicMock
+from collections.abc import AsyncIterator
+from concurrent.futures import ThreadPoolExecutor
+from contextlib import asynccontextmanager
+from typing import Any, Optional

 import pytest
-from conftest import DEFAULT_URI  # type: ignore
+from langchain_core.embeddings import Embeddings
+from psycopg import AsyncConnection

-from langgraph.store.base import GetOp, Item, ListNamespacesOp, PutOp, SearchOp
+from langgraph.store.base import (
+    GetOp,
+    Item,
+    ListNamespacesOp,
+    PutOp,
+    SearchOp,
+)
 from langgraph.store.postgres import AsyncPostgresStore
+from tests.conftest import (
+    DEFAULT_URI,
+    VECTOR_TYPES,
+    CharacterEmbeddings,
+)


-class MockAsyncCursor:
-    def __init__(self, fetch_result: Any) -> None:
-        self.fetch_result = fetch_result
-        self.execute = AsyncMock()
-        self.fetchall = AsyncMock(return_value=self.fetch_result)
+@pytest.fixture(scope="function", params=["default", "pipe", "pool"])
+async def store(request) -> AsyncIterator[AsyncPostgresStore]:
+    if sys.version_info < (3, 10):
+        pytest.skip("Async Postgres tests require Python 3.10+")
+
+    database = f"test_{uuid.uuid4().hex[:16]}"
+    uri_parts = DEFAULT_URI.split("/")
+    uri_base = "/".join(uri_parts[:-1])
+    query_params = ""
+    if "?" in uri_parts[-1]:
+        db_name, query_params = uri_parts[-1].split("?", 1)
+        query_params = "?" + query_params
+
+    conn_string = f"{uri_base}/{database}{query_params}"
+    admin_conn_string = DEFAULT_URI
+
+    async with await AsyncConnection.connect(
+        admin_conn_string, autocommit=True
+    ) as conn:
+        await conn.execute(f"CREATE DATABASE {database}")
+    try:
+        async with AsyncPostgresStore.from_conn_string(conn_string) as store:
+            await store.setup()
+
+        if request.param == "pipe":
+            async with AsyncPostgresStore.from_conn_string(
+                conn_string, pipeline=True
+            ) as store:
+                yield store
+        elif request.param == "pool":
+            async with AsyncPostgresStore.from_conn_string(
+                conn_string, pool_config={"min_size": 1, "max_size": 10}
+            ) as store:
+                yield store
+        else:  # default
+            async with AsyncPostgresStore.from_conn_string(conn_string) as store:
+                yield store
+    finally:
+        async with await AsyncConnection.connect(
+            admin_conn_string, autocommit=True
+        ) as conn:
+            await conn.execute(f"DROP DATABASE {database}")


-class MockAsyncConnection:
-    def __init__(self) -> None:
-        self.cursor = MagicMock()
-        self.pipeline = MagicMock(
-            return_value=AsyncMock(__aenter__=AsyncMock(), __aexit__=AsyncMock())
+async def test_no_running_loop(store: AsyncPostgresStore) -> None:
+    with pytest.raises(asyncio.InvalidStateError):
+        store.put(("foo", "bar"), "baz", {"val": "baz"})
+    with pytest.raises(asyncio.InvalidStateError):
+        store.get(("foo", "bar"), "baz")
+    with pytest.raises(asyncio.InvalidStateError):
+        store.delete(("foo", "bar"), "baz")
+    with pytest.raises(asyncio.InvalidStateError):
+        store.search(("foo", "bar"))
+    with pytest.raises(asyncio.InvalidStateError):
+        store.list_namespaces(prefix=("foo",))
+    with pytest.raises(asyncio.InvalidStateError):
+        store.batch([PutOp(namespace=("foo", "bar"), key="baz", value={"val": "baz"})])
+    with ThreadPoolExecutor(max_workers=1) as executor:
+        future = executor.submit(store.put, ("foo", "bar"), "baz", {"val": "baz"})
+        result = await asyncio.wrap_future(future)
+        assert result is None
+        future = executor.submit(store.get, ("foo", "bar"), "baz")
+        result = await asyncio.wrap_future(future)
+        assert result.value == {"val": "baz"}
+        result = await asyncio.wrap_future(
+            executor.submit(store.list_namespaces, prefix=("foo",))
        )


-@pytest.fixture
-def mock_connection() -> MockAsyncConnection:
-    return MockAsyncConnection()
+async def test_large_batches(request: Any, store: AsyncPostgresStore) -> None:
+    N = 100  # less important that we are performant here
+    M = 10
+
+    with ThreadPoolExecutor(max_workers=10) as executor:
+        futures = []
+        for m in range(M):
+            for i in range(N):
+                futures += [
+                    executor.submit(
+                        store.put,
+                        ("test", "foo", "bar", "baz", str(m % 2)),
+                        f"key{i}",
+                        value={"foo": "bar" + str(i)},
+                    ),
+                    executor.submit(
+                        store.get,
+                        ("test", "foo", "bar", "baz", str(m % 2)),
+                        f"key{i}",
+                    ),
+                    executor.submit(
+                        store.list_namespaces,
+                        prefix=None,
+                        max_depth=m + 1,
+                    ),
+                    executor.submit(
+                        store.search,
+                        ("test",),
+                    ),
+                    executor.submit(
+                        store.put,
+                        ("test", "foo", "bar", "baz", str(m % 2)),
+                        f"key{i}",
+                        value={"foo": "bar" + str(i)},
+                    ),
+                    executor.submit(
+                        store.put,
+                        ("test", "foo", "bar", "baz", str(m % 2)),
+                        f"key{i}",
+                        None,
+                    ),
+                ]
+
+        results = await asyncio.gather(
+            *(asyncio.wrap_future(future) for future in futures)
+        )
+    assert len(results) == M * N * 6


-@pytest.fixture
-async def store(mock_connection: MockAsyncConnection) -> AsyncPostgresStore:
-    return AsyncPostgresStore(mock_connection)
+async def test_large_batches_async(store: AsyncPostgresStore) -> None:
+    N = 1000
+    M = 10
+    coros = []
+    for m in range(M):
+        for i in range(N):
+            coros.append(
+                store.aput(
+                    ("test", "foo", "bar", "baz", str(m % 2)),
+                    f"key{i}",
+                    value={"foo": "bar" + str(i)},
+                )
+            )
+            coros.append(
+                store.aget(
+                    ("test", "foo", "bar", "baz", str(m % 2)),
+                    f"key{i}",
+                )
+            )
+            coros.append(
+                store.alist_namespaces(
+                    prefix=None,
+                    max_depth=m + 1,
+                )
+            )
+            coros.append(
+                store.asearch(
+                    ("test",),
+                )
+            )
+            coros.append(
+                store.aput(
+                    ("test", "foo", "bar", "baz", str(m % 2)),
+                    f"key{i}",
+                    value={"foo": "bar" + str(i)},
+                )
+            )
+            coros.append(
+                store.adelete(
+                    ("test", "foo", "bar", "baz", str(m % 2)),
+                    f"key{i}",
+                )
+            )
+
+    results = await asyncio.gather(*coros)
+    assert len(results) == M * N * 6


 async def test_abatch_order(store: AsyncPostgresStore) -> None:
-    mock_connection = store.conn
-    mock_get_cursor = MockAsyncCursor(
-        [
-            {
-                "key": "key1",
-                "value": '{"data": "value1"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.foo",
-            },
-            {
-                "key": "key2",
-                "value": '{"data": "value2"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.bar",
-            },
-        ]
-    )
-    mock_search_cursor = MockAsyncCursor(
-        [
-            {
-                "key": "key1",
-                "value": '{"data": "value1"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.foo",
-            },
-        ]
-    )
-    mock_list_namespaces_cursor = MockAsyncCursor(
-        [
-            {"truncated_prefix": b"\x01test"},
-        ]
-    )
-
-    failures = []
-
-    def cursor_side_effect(binary: bool = False) -> Any:
-        cursor = MagicMock()
-
-        async def execute_side_effect(query: str, *params: Any) -> None:
-            # My super sophisticated database.
-            if "SELECT prefix, key," in query:
-                cursor.fetchall = mock_search_cursor.fetchall
-            elif "SELECT DISTINCT ON (truncated_prefix)" in query:
-                cursor.fetchall = mock_list_namespaces_cursor.fetchall
-            elif "WHERE prefix = %s AND key" in query:
-                cursor.fetchall = mock_get_cursor.fetchall
-            elif "INSERT INTO " in query:
-                pass
-            else:
-                e = ValueError(f"Unmatched query: {query}")
-                failures.append(e)
-                raise e
-
-        cursor.execute = AsyncMock(side_effect=execute_side_effect)
-        return cursor
-
-    mock_connection.cursor.side_effect = cursor_side_effect  # type: ignore
+    # Setup test data
+    await store.aput(("test", "foo"), "key1", {"data": "value1"})
+    await store.aput(("test", "bar"), "key2", {"data": "value2"})

    ops = [
-        GetOp(namespace=("test",), key="key1"),
-        PutOp(namespace=("test",), key="key2", value={"data": "value2"}),
+        GetOp(namespace=("test", "foo"), key="key1"),
+        PutOp(namespace=("test", "bar"), key="key2", value={"data": "value2"}),
        SearchOp(
            namespace_prefix=("test",), filter={"data": "value1"}, limit=10, offset=0
        ),
        ListNamespacesOp(match_conditions=None, max_depth=None, limit=10, offset=0),
        GetOp(namespace=("test",), key="key3"),
    ]
+
    results = await store.abatch(ops)
-    assert not failures
    assert len(results) == 5
    assert isinstance(results[0], Item)
    assert isinstance(results[0].value, dict)
@@ -118,27 +218,29 @@ async def test_abatch_order(store: AsyncPostgresStore) -> None:
    assert isinstance(results[2], list)
    assert len(results[2]) == 1
    assert isinstance(results[3], list)
-    assert results[3] == [("test",)]
+    assert ("test", "foo") in results[3] and ("test", "bar") in results[3]
    assert results[4] is None

    ops_reordered = [
        SearchOp(namespace_prefix=("test",), filter=None, limit=5, offset=0),
-        GetOp(namespace=("test",), key="key2"),
+        GetOp(namespace=("test", "bar"), key="key2"),
        ListNamespacesOp(match_conditions=None, max_depth=None, limit=5, offset=0),
        PutOp(namespace=("test",), key="key3", value={"data": "value3"}),
-        GetOp(namespace=("test",), key="key1"),
+        GetOp(namespace=("test", "foo"), key="key1"),
    ]

    results_reordered = await store.abatch(ops_reordered)
-    assert not failures
    assert len(results_reordered) == 5
    assert isinstance(results_reordered[0], list)
-    assert len(results_reordered[0]) == 1
+    assert len(results_reordered[0]) == 2
    assert isinstance(results_reordered[1], Item)
    assert results_reordered[1].value == {"data": "value2"}
    assert results_reordered[1].key == "key2"
    assert isinstance(results_reordered[2], list)
-    assert results_reordered[2] == [("test",)]
+    assert ("test", "foo") in results_reordered[2] and (
+        "test",
+        "bar",
+    ) in results_reordered[2]
    assert results_reordered[3] is None
    assert isinstance(results_reordered[4], Item)
    assert results_reordered[4].value == {"data": "value1"}
@@ -146,26 +248,9 @@ async def test_abatch_order(store: AsyncPostgresStore) -> None:


 async def test_batch_get_ops(store: AsyncPostgresStore) -> None:
-    mock_connection = store.conn
-    mock_cursor = MockAsyncCursor(
-        [
-            {
-                "key": "key1",
-                "value": '{"data": "value1"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.foo",
-            },
-            {
-                "key": "key2",
-                "value": '{"data": "value2"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.bar",
-            },
-        ]
-    )
-    mock_connection.cursor.return_value = mock_cursor
+    # Setup test data
+    await store.aput(("test",), "key1", {"data": "value1"})
+    await store.aput(("test",), "key2", {"data": "value2"})

    ops = [
        GetOp(namespace=("test",), key="key1"),
@@ -184,10 +269,6 @@ async def test_batch_get_ops(store: AsyncPostgresStore) -> None:


 async def test_batch_put_ops(store: AsyncPostgresStore) -> None:
-    mock_connection = store.conn
-    mock_cursor = MockAsyncCursor([])
-    mock_connection.cursor.return_value = mock_cursor
-
    ops = [
        PutOp(namespace=("test",), key="key1", value={"data": "value1"}),
        PutOp(namespace=("test",), key="key2", value={"data": "value2"}),
@@ -198,30 +279,16 @@ async def test_batch_put_ops(store: AsyncPostgresStore) -> None:

    assert len(results) == 3
    assert all(result is None for result in results)
-    assert mock_cursor.execute.call_count == 2
+
+    # Verify the puts worked
+    items = await store.asearch(["test"], limit=10)
+    assert len(items) == 2  # key3 had None value so wasn't stored


 async def test_batch_search_ops(store: AsyncPostgresStore) -> None:
-    mock_connection = store.conn
-    mock_cursor = MockAsyncCursor(
-        [
-            {
-                "key": "key1",
-                "value": '{"data": "value1"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.foo",
-            },
-            {
-                "key": "key2",
-                "value": '{"data": "value2"}',
-                "created_at": datetime.now(),
-                "updated_at": datetime.now(),
-                "prefix": "test.bar",
-            },
-        ]
-    )
-    mock_connection.cursor.return_value = mock_cursor
+    # Setup test data
+    await store.aput(("test", "foo"), "key1", {"data": "value1"})
+    await store.aput(("test", "bar"), "key2", {"data": "value2"})

    ops = [
        SearchOp(
@@ -233,297 +300,338 @@ async def test_batch_search_ops(store: AsyncPostgresStore) -> None:
    results = await store.abatch(ops)

    assert len(results) == 2
-    assert len(results[0]) == 2
-    assert len(results[1]) == 2
+    assert len(results[0]) == 1  # Filtered results
+    assert len(results[1]) == 2  # All results


 async def test_batch_list_namespaces_ops(store: AsyncPostgresStore) -> None:
-    mock_connection = store.conn
-    mock_cursor = MockAsyncCursor(
-        [
-            {"truncated_prefix": b"\x01test.namespace1"},
-            {"truncated_prefix": b"\x01test.namespace2"},
-        ]
-    )
-    mock_connection.cursor.return_value = mock_cursor
+    # Setup test data
+    await store.aput(("test", "namespace1"), "key1", {"data": "value1"})
+    await store.aput(("test", "namespace2"), "key2", {"data": "value2"})

    ops = [ListNamespacesOp(match_conditions=None, max_depth=None, limit=10, offset=0)]

    results = await store.abatch(ops)

    assert len(results) == 1
-    assert results[0] == [("test", "namespace1"), ("test", "namespace2")]
+    assert len(results[0]) == 2
+    assert ("test", "namespace1") in results[0]
+    assert ("test", "namespace2") in results[0]


-# The following use the actual DB connection
+@asynccontextmanager
+async def _create_vector_store(
+    vector_type: str,
+    distance_type: str,
+    fake_embeddings: CharacterEmbeddings,
+    text_fields: Optional[list[str]] = None,
+) -> AsyncIterator[AsyncPostgresStore]:
+    """Create a store with vector search enabled."""
+    if sys.version_info < (3, 10):
+        pytest.skip("Async Postgres tests require Python 3.10+")

+    database = f"test_{uuid.uuid4().hex[:16]}"
+    uri_parts = DEFAULT_URI.split("/")
+    uri_base = "/".join(uri_parts[:-1])
+    query_params = ""
+    if "?" in uri_parts[-1]:
+        db_name, query_params = uri_parts[-1].split("?", 1)
+        query_params = "?" + query_params

-class TestAsyncPostgresStore:
-    @pytest.fixture(autouse=True)
-    async def setup(self) -> None:
-        async with AsyncPostgresStore.from_conn_string(DEFAULT_URI) as store:
+    conn_string = f"{uri_base}/{database}{query_params}"
+    admin_conn_string = DEFAULT_URI
+
+    index_config = {
+        "dims": fake_embeddings.dims,
+        "embed": fake_embeddings,
+        "ann_index_config": {
+            "vector_type": vector_type,
+        },
+        "distance_type": distance_type,
+        "text_fields": text_fields,
+    }
+
+    async with await AsyncConnection.connect(
+        admin_conn_string, autocommit=True
+    ) as conn:
+        await conn.execute(f"CREATE DATABASE {database}")
+    try:
+        async with AsyncPostgresStore.from_conn_string(
+            conn_string,
+            index=index_config,
+        ) as store:
            await store.setup()
+            yield store
+    finally:
+        async with await AsyncConnection.connect(
+            admin_conn_string, autocommit=True
+        ) as conn:
+            await conn.execute(f"DROP DATABASE {database}")

-    async def test_basic_store_ops(self) -> None:
-        async with AsyncPostgresStore.from_conn_string(DEFAULT_URI) as store:
-            namespace = ("test", "documents")
-            item_id = "doc1"
-            item_value = {"title": "Test Document", "content": "Hello, World!"}

-            await store.aput(namespace, item_id, item_value)
-            item = await store.aget(namespace, item_id)
+@pytest.fixture(
+    scope="function",
+    params=[
+        (vector_type, distance_type)
+        for vector_type in VECTOR_TYPES
+        for distance_type in (
+            ["hamming"] if vector_type == "bit" else ["l2", "inner_product", "cosine"]
+        )
+    ],
+    ids=lambda p: f"{p[0]}_{p[1]}",
+)
+async def vector_store(
+    request,
+    fake_embeddings: CharacterEmbeddings,
+) -> AsyncIterator[AsyncPostgresStore]:
+    """Create a store with vector search enabled."""
+    vector_type, distance_type = request.param
+    async with _create_vector_store(
+        vector_type, distance_type, fake_embeddings
+    ) as store:
+        yield store

-            assert item
-            assert item.namespace == namespace
-            assert item.key == item_id
-            assert item.value == item_value

-            updated_value = {
-                "title": "Updated Test Document",
-                "content": "Hello, LangGraph!",
-            }
-            await store.aput(namespace, item_id, updated_value)
-            updated_item = await store.aget(namespace, item_id)
+async def test_vector_store_initialization(
+    vector_store: AsyncPostgresStore, fake_embeddings: CharacterEmbeddings
+) -> None:
+    """Test store initialization with embedding config."""
+    assert vector_store.index_config is not None
+    assert vector_store.index_config["dims"] == fake_embeddings.dims
+    if isinstance(vector_store.index_config["embed"], Embeddings):
+        assert vector_store.index_config["embed"] == fake_embeddings

-            assert updated_item.value == updated_value
-            assert updated_item.updated_at > item.updated_at
-            different_namespace = ("test", "other_documents")
-            item_in_different_namespace = await store.aget(different_namespace, item_id)
-            assert item_in_different_namespace is None

-            new_item_id = "doc2"
-            new_item_value = {"title": "Another Document", "content": "Greetings!"}
-            await store.aput(namespace, new_item_id, new_item_value)
+async def test_vector_insert_with_auto_embedding(
+    vector_store: AsyncPostgresStore,
+) -> None:
+    """Test inserting items that get auto-embedded."""
+    docs = [
+        ("doc1", {"text": "short text"}),
+        ("doc2", {"text": "longer text document"}),
+        ("doc3", {"text": "longest text document here"}),
+        ("doc4", {"description": "text in description field"}),
+        ("doc5", {"content": "text in content field"}),
+        ("doc6", {"body": "text in body field"}),
+    ]

-            search_results = await store.asearch(["test"], limit=10)
-            items = search_results
-            assert len(items) == 2
-            assert any(item.key == item_id for item in items)
-            assert any(item.key == new_item_id for item in items)
+    for key, value in docs:
+        await vector_store.aput(("test",), key, value)

-            namespaces = await store.alist_namespaces(prefix=["test"])
-            assert ("test", "documents") in namespaces
+    results = await vector_store.asearch(("test",), query="long text")
+    assert len(results) > 0

-            await store.adelete(namespace, item_id)
-            await store.adelete(namespace, new_item_id)
-            deleted_item = await store.aget(namespace, item_id)
-            assert deleted_item is None
+    doc_order = [r.key for r in results]
+    assert "doc2" in doc_order
+    assert "doc3" in doc_order

-            deleted_item = await store.aget(namespace, new_item_id)
-            assert deleted_item is None

-            empty_search_results = await store.asearch(["test"], limit=10)
-            assert len(empty_search_results) == 0
+async def test_vector_update_with_embedding(vector_store: AsyncPostgresStore) -> None:
+    """Test that updating items properly updates their embeddings."""
+    await vector_store.aput(("test",), "doc1", {"text": "zany zebra Xerxes"})
+    await vector_store.aput(("test",), "doc2", {"text": "something about dogs"})
+    await vector_store.aput(("test",), "doc3", {"text": "text about birds"})

-    async def test_list_namespaces(self) -> None:
-        async with AsyncPostgresStore.from_conn_string(DEFAULT_URI) as store:
-            test_pref = str(uuid.uuid4())
-            test_namespaces = [
-                (test_pref, "test", "documents", "public", test_pref),
-                (test_pref, "test", "documents", "private", test_pref),
-                (test_pref, "test", "images", "public", test_pref),
-                (test_pref, "test", "images", "private", test_pref),
-                (test_pref, "prod", "documents", "public", test_pref),
-                (
-                    test_pref,
-                    "prod",
-                    "documents",
-                    "some",
-                    "nesting",
-                    "public",
-                    test_pref,
-                ),
-                (test_pref, "prod", "documents", "private", test_pref),
-            ]
+    results_initial = await vector_store.asearch(("test",), query="Zany Xerxes")
+    assert len(results_initial) > 0
+    assert results_initial[0].key == "doc1"
+    initial_score = results_initial[0].score

-            for namespace in test_namespaces:
-                await store.aput(namespace, "dummy", {"content": "dummy"})
+    await vector_store.aput(("test",), "doc1", {"text": "new text about dogs"})

-            prefix_result = await store.alist_namespaces(prefix=[test_pref, "test"])
-            assert len(prefix_result) == 4
-            assert all([ns[1] == "test" for ns in prefix_result])
+    results_after = await vector_store.asearch(("test",), query="Zany Xerxes")
+    after_score = next((r.score for r in results_after if r.key == "doc1"), 0.0)
+    assert after_score < initial_score

-            specific_prefix_result = await store.alist_namespaces(
-                prefix=[test_pref, "test", "documents"]
-            )
-            assert len(specific_prefix_result) == 2
-            assert all(
-                [ns[1:3] == ("test", "documents") for ns in specific_prefix_result]
-            )
+    results_new = await vector_store.asearch(("test",), query="new text about dogs")
+    for r in results_new:
+        if r.key == "doc1":
+            assert r.score > after_score

-            suffix_result = await store.alist_namespaces(suffix=["public", test_pref])
-            assert len(suffix_result) == 4
-            assert all(ns[-2] == "public" for ns in suffix_result)
+    # Don't index this one
+    await vector_store.aput(
+        ("test",), "doc4", {"text": "new text about dogs"}, index=False
+    )
+    results_new = await vector_store.asearch(
+        ("test",), query="new text about dogs", limit=3
+    )
+    assert not any(r.key == "doc4" for r in results_new)

-            prefix_suffix_result = await store.alist_namespaces(
-                prefix=[test_pref, "test"], suffix=["public", test_pref]
-            )
-            assert len(prefix_suffix_result) == 2
-            assert all(
-                ns[1] == "test" and ns[-2] == "public" for ns in prefix_suffix_result
-            )

-            wildcard_prefix_result = await store.alist_namespaces(
-                prefix=[test_pref, "*", "documents"]
-            )
-            assert len(wildcard_prefix_result) == 5
-            assert all(ns[2] == "documents" for ns in wildcard_prefix_result)
+async def test_vector_search_with_filters(vector_store: AsyncPostgresStore) -> None:
+    """Test combining vector search with filters."""
+    docs = [
+        ("doc1", {"text": "red apple", "color": "red", "score": 4.5}),
+        ("doc2", {"text": "red car", "color": "red", "score": 3.0}),
+        ("doc3", {"text": "green apple", "color": "green", "score": 4.0}),
+        ("doc4", {"text": "blue car", "color": "blue", "score": 3.5}),
+    ]

-            wildcard_suffix_result = await store.alist_namespaces(
-                suffix=["*", "public", test_pref]
-            )
-            assert len(wildcard_suffix_result) == 4
-            assert all(ns[-2] == "public" for ns in wildcard_suffix_result)
-            wildcard_single = await store.alist_namespaces(
-                suffix=["some", "*", "public", test_pref]
-            )
-            assert len(wildcard_single) == 1
-            assert wildcard_single[0] == (
-                test_pref,
-                "prod",
-                "documents",
-                "some",
-                "nesting",
-                "public",
-                test_pref,
-            )
+    for key, value in docs:
+        await vector_store.aput(("test",), key, value)

-            max_depth_result = await store.alist_namespaces(max_depth=3)
-            assert all([len(ns) <= 3 for ns in max_depth_result])
-            max_depth_result = await store.alist_namespaces(
-                max_depth=4, prefix=[test_pref, "*", "documents"]
-            )
-            assert (
-                len(set(tuple(res) for res in max_depth_result))
-                == len(max_depth_result)
-                == 5
-            )
+    results = await vector_store.asearch(
+        ("test",), query="apple", filter={"color": "red"}
+    )
+    assert len(results) == 2
+    assert results[0].key == "doc1"

-            limit_result = await store.alist_namespaces(prefix=[test_pref], limit=3)
-            assert len(limit_result) == 3
+    results = await vector_store.asearch(
+        ("test",), query="car", filter={"color": "red"}
+    )
+    assert len(results) == 2
+    assert results[0].key == "doc2"

-            offset_result = await store.alist_namespaces(prefix=[test_pref], offset=3)
-            assert len(offset_result) == len(test_namespaces) - 3
+    results = await vector_store.asearch(
+        ("test",), query="bbbbluuu", filter={"score": {"$gt": 3.2}}
+    )
+    assert len(results) == 3
+    assert results[0].key == "doc4"

-            empty_prefix_result = await store.alist_namespaces(prefix=[test_pref])
-            assert len(empty_prefix_result) == len(test_namespaces)
-            assert set(tuple(ns) for ns in empty_prefix_result) == set(
-                tuple(ns) for ns in test_namespaces
-            )
+    results = await vector_store.asearch(
+        ("test",), query="apple", filter={"score": {"$gte": 4.0}, "color": "green"}
+    )
+    assert len(results) == 1
+    assert results[0].key == "doc3"

-            for namespace in test_namespaces:
-                await store.adelete(namespace, "dummy")

-    async def test_search(self):
-        async with AsyncPostgresStore.from_conn_string(DEFAULT_URI) as store:
-            test_namespaces = [
-                ("test_search", "documents", "user1"),
-                ("test_search", "documents", "user2"),
-                ("test_search", "reports", "department1"),
-                ("test_search", "reports", "department2"),
-            ]
-            test_items = [
-                {"title": "Doc 1", "author": "John Doe", "tags": ["important"]},
-                {"title": "Doc 2", "author": "Jane Smith", "tags": ["draft"]},
-                {"title": "Report A", "author": "John Doe", "tags": ["final"]},
-                {"title": "Report B", "author": "Alice Johnson", "tags": ["draft"]},
-            ]
-            empty = await store.asearch(
-                (
-                    "scoped",
-                    "assistant_id",
-                    "shared",
-                    "6c5356f6-63ab-4158-868d-cd9fd14c736e",
-                ),
-                limit=10,
-                offset=0,
-            )
-            assert len(empty) == 0
+async def test_vector_search_pagination(vector_store: AsyncPostgresStore) -> None:
+    """Test pagination with vector search."""
+    for i in range(5):
+        await vector_store.aput(
+            ("test",), f"doc{i}", {"text": f"test document number {i}"}
+        )

-            for namespace, item in zip(test_namespaces, test_items):
-                await store.aput(namespace, f"item_{namespace[-1]}", item)
+    results_page1 = await vector_store.asearch(("test",), query="test", limit=2)
+    results_page2 = await vector_store.asearch(
+        ("test",), query="test", limit=2, offset=2
+    )

-            docs_result = await store.asearch(["test_search", "documents"])
-            assert len(docs_result) == 2
-            assert all([item.namespace[1] == "documents" for item in docs_result]), [
-                item.namespace for item in docs_result
-            ]
+    assert len(results_page1) == 2
+    assert len(results_page2) == 2
+    assert results_page1[0].key != results_page2[0].key

-            reports_result = await store.asearch(["test_search", "reports"])
-            assert len(reports_result) == 2
-            assert all(item.namespace[1] == "reports" for item in reports_result)
+    all_results = await vector_store.asearch(("test",), query="test", limit=10)
+    assert len(all_results) == 5

-            limited_result = await store.asearch(["test_search"], limit=2)
-            assert len(limited_result) == 2
-            offset_result = await store.asearch(["test_search"])
-            assert len(offset_result) == 4

-            offset_result = await store.asearch(["test_search"], offset=2)
-            assert len(offset_result) == 2
-            assert all(item not in limited_result for item in offset_result)
+async def test_vector_search_edge_cases(vector_store: AsyncPostgresStore) -> None:
+    """Test edge cases in vector search."""
+    await vector_store.aput(("test",), "doc1", {"text": "test document"})

-            john_doe_result = await store.asearch(
-                ["test_search"], filter={"author": "John Doe"}
-            )
-            assert len(john_doe_result) == 2
-            assert all(item.value["author"] == "John Doe" for item in john_doe_result)
+    perfect_match = await vector_store.asearch(("test",), query="text test document")
+    perfect_score = perfect_match[0].score

-            draft_result = await store.asearch(
-                ["test_search"], filter={"tags": ["draft"]}
-            )
-            assert len(draft_result) == 2
-            assert all("draft" in item.value["tags"] for item in draft_result)
+    results = await vector_store.asearch(("test",), query="")
+    assert len(results) == 1
+    assert results[0].score is None

-            page1 = await store.asearch(["test_search"], limit=2, offset=0)
-            page2 = await store.asearch(["test_search"], limit=2, offset=2)
-            all_items = page1 + page2
-            assert len(all_items) == 4
-            assert len(set(item.key for item in all_items)) == 4
-            empty = await store.asearch(
-                (
-                    "scoped",
-                    "assistant_id",
-                    "shared",
-                    "again",
-                    "maybe",
-                    "some-long",
-                    "6be5cb0e-2eb4-42e6-bb6b-fba3c269db25",
-                ),
-                limit=10,
-                offset=0,
-            )
-            assert len(empty) == 0
+    results = await vector_store.asearch(("test",), query=None)
+    assert len(results) == 1
+    assert results[0].score is None

-            # Test with a namespace beginning with a number (like a UUID)
-            uuid_namespace = (str(uuid.uuid4()), "documents")
-            uuid_item_id = "uuid_doc"
-            uuid_item_value = {
-                "title": "UUID Document",
-                "content": "This document has a UUID namespace.",
-            }
+    long_query = "foo " * 100
+    results = await vector_store.asearch(("test",), query=long_query)
+    assert len(results) == 1
+    assert results[0].score < perfect_score

-            # Insert the item with the UUID namespace
-            await store.aput(uuid_namespace, uuid_item_id, uuid_item_value)
+    special_query = "test!@#$%^&*()"
+    results = await vector_store.asearch(("test",), query=special_query)
+    assert len(results) == 1
+    assert results[0].score < perfect_score

-            # Retrieve the item to verify it was stored correctly
-            retrieved_item = await store.aget(uuid_namespace, uuid_item_id)
-            assert retrieved_item is not None
-            assert retrieved_item.namespace == uuid_namespace
-            assert retrieved_item.key == uuid_item_id
-            assert retrieved_item.value == uuid_item_value

-            # Search for the item using the UUID namespace
-            search_result = await store.asearch([uuid_namespace[0]])
-            assert len(search_result) == 1
-            assert search_result[0].key == uuid_item_id
-            assert search_result[0].value == uuid_item_value
+@pytest.mark.parametrize(
+    "vector_type,distance_type",
+    [
+        *itertools.product(["vector", "halfvec"], ["cosine", "inner_product", "l2"]),
+    ],
+)
+async def test_embed_with_path(
+    request: Any,
+    fake_embeddings: CharacterEmbeddings,
+    vector_type: str,
+    distance_type: str,
+) -> None:
+    """Test vector search with specific text fields in Postgres store."""
+    async with _create_vector_store(
+        vector_type,
+        distance_type,
+        fake_embeddings,
+        text_fields=["key0", "key1", "key3"],
+    ) as store:
+        # This will have 2 vectors representing it
+        doc1 = {
+            # Omit key0 - check it doesn't raise an error
+            "key1": "xxx",
+            "key2": "yyy",
+            "key3": "zzz",
+        }
+        # This will have 3 vectors representing it
+        doc2 = {
+            "key0": "uuu",
+            "key1": "vvv",
+            "key2": "www",
+            "key3": "xxx",
+        }
+        await store.aput(("test",), "doc1", doc1)
+        await store.aput(("test",), "doc2", doc2)

-            # Clean up: delete the item with the UUID namespace
-            await store.adelete(uuid_namespace, uuid_item_id)
+        # doc2.key3 and doc1.key1 both would have the highest score
+        results = await store.asearch(("test",), query="xxx")
+        assert len(results) == 2
+        assert results[0].key != results[1].key
+        ascore = results[0].score
+        bscore = results[1].score
+        assert ascore == pytest.approx(bscore, abs=1e-3)

-            # Verify the item was deleted
-            deleted_item = await store.aget(uuid_namespace, uuid_item_id)
-            assert deleted_item is None
+        results = await store.asearch(("test",), query="uuu")
+        assert len(results) == 2
+        assert results[0].key != results[1].key
+        assert results[0].key == "doc2"
+        assert results[0].score > results[1].score
+        assert ascore == pytest.approx(results[0].score, abs=1e-3)

-            for namespace in test_namespaces:
-                await store.adelete(namespace, f"item_{namespace[-1]}")
+        # Un-indexed - will have low results for both. Not zero (because we're projecting)
+        # but less than the above.
+        results = await store.asearch(("test",), query="www")
+        assert len(results) == 2
+        assert results[0].score < ascore
+        assert results[1].score < ascore
+
+
+@pytest.mark.parametrize(
+    "vector_type,distance_type",
+    [
+        *itertools.product(["vector", "halfvec"], ["cosine", "inner_product", "l2"]),
+    ],
+)
+async def test_search_sorting(
+    request: Any,
+    fake_embeddings: CharacterEmbeddings,
+    vector_type: str,
+    distance_type: str,
+) -> None:
+    """Test operation-level field configuration for vector search."""
+    async with _create_vector_store(
+        vector_type,
+        distance_type,
+        fake_embeddings,
+        text_fields=["key1"],  # Default fields that won't match our test data
+    ) as store:
+        amatch = {
+            "key1": "mmm",
+        }
+
+        await store.aput(("test", "M"), "M", amatch)
+        N = 100
+        for i in range(N):
+            await store.aput(("test", "A"), f"A{i}", {"key1": "no"})
+        for i in range(N):
+            await store.aput(("test", "Z"), f"Z{i}", {"key1": "no"})
+
+        results = await store.asearch(("test",), query="mmm", limit=10)
+        assert len(results) == 10
+        assert len(set(r.key for r in results)) == 10
+        assert results[0].key == "M"
+        assert results[0].score > results[1].score
@@ -1,8 +1,14 @@
+# type: ignore
+
+from contextlib import contextmanager
 from typing import Any
+from uuid import uuid4

 import pytest
-from conftest import DEFAULT_URI  # type: ignore
 from langchain_core.runnables import RunnableConfig
+from psycopg import Connection
+from psycopg.rows import dict_row
+from psycopg_pool import ConnectionPool

 from langgraph.checkpoint.base import (
    Checkpoint,
@@ -11,102 +17,198 @@ from langgraph.checkpoint.base import (
    empty_checkpoint,
 )
 from langgraph.checkpoint.postgres import PostgresSaver
+from tests.conftest import DEFAULT_POSTGRES_URI


-class TestPostgresSaver:
-    @pytest.fixture(autouse=True)
-    def setup(self) -> None:
-        # objects for test setup
-        self.config_1: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-1",
-                # for backwards compatibility testing
-                "thread_ts": "1",
-                "checkpoint_ns": "",
-            }
-        }
-        self.config_2: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-2",
-                "checkpoint_id": "2",
-                "checkpoint_ns": "",
-            }
-        }
-        self.config_3: RunnableConfig = {
-            "configurable": {
-                "thread_id": "thread-2",
-                "checkpoint_id": "2-inner",
-                "checkpoint_ns": "inner",
-            }
-        }
+@contextmanager
+def _pool_saver():
+    """Fixture for pool mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+        conn.execute(f"CREATE DATABASE {database}")
+    try:
+        # yield checkpointer
+        with ConnectionPool(
+            DEFAULT_POSTGRES_URI + database,
+            max_size=10,
+            kwargs={"autocommit": True, "row_factory": dict_row},
+        ) as pool:
+            checkpointer = PostgresSaver(pool)
+            checkpointer.setup()
+            yield checkpointer
+    finally:
+        # drop unique db
+        with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+            conn.execute(f"DROP DATABASE {database}")

-        self.chkpnt_1: Checkpoint = empty_checkpoint()
-        self.chkpnt_2: Checkpoint = create_checkpoint(self.chkpnt_1, {}, 1)
-        self.chkpnt_3: Checkpoint = empty_checkpoint()

-        self.metadata_1: CheckpointMetadata = {
-            "source": "input",
-            "step": 2,
-            "writes": {},
-            "score": 1,
+@contextmanager
+def _pipe_saver():
+    """Fixture for pipeline mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+        conn.execute(f"CREATE DATABASE {database}")
+    try:
+        with Connection.connect(
+            DEFAULT_POSTGRES_URI + database,
+            autocommit=True,
+            prepare_threshold=0,
+            row_factory=dict_row,
+        ) as conn:
+            checkpointer = PostgresSaver(conn)
+            checkpointer.setup()
+            with conn.pipeline() as pipe:
+                checkpointer = PostgresSaver(conn, pipe=pipe)
+                yield checkpointer
+    finally:
+        # drop unique db
+        with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+            conn.execute(f"DROP DATABASE {database}")
+
+
+@contextmanager
+def _base_saver():
+    """Fixture for regular connection mode testing."""
+    database = f"test_{uuid4().hex[:16]}"
+    # create unique db
+    with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+        conn.execute(f"CREATE DATABASE {database}")
+    try:
+        with Connection.connect(
+            DEFAULT_POSTGRES_URI + database,
+            autocommit=True,
+            prepare_threshold=0,
+            row_factory=dict_row,
+        ) as conn:
+            checkpointer = PostgresSaver(conn)
+            checkpointer.setup()
+            yield checkpointer
+    finally:
+        # drop unique db
+        with Connection.connect(DEFAULT_POSTGRES_URI, autocommit=True) as conn:
+            conn.execute(f"DROP DATABASE {database}")
+
+
+@contextmanager
+def _saver(name: str):
+    if name == "base":
+        with _base_saver() as saver:
+            yield saver
+    elif name == "pool":
+        with _pool_saver() as saver:
+            yield saver
+    elif name == "pipe":
+        with _pipe_saver() as saver:
+            yield saver
+
+
+@pytest.fixture
+def test_data():
+    """Fixture providing test data for checkpoint tests."""
+    config_1: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-1",
+            # for backwards compatibility testing
+            "thread_ts": "1",
+            "checkpoint_ns": "",
        }
-        self.metadata_2: CheckpointMetadata = {
-            "source": "loop",
+    }
+    config_2: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-2",
+            "checkpoint_id": "2",
+            "checkpoint_ns": "",
+        }
+    }
+    config_3: RunnableConfig = {
+        "configurable": {
+            "thread_id": "thread-2",
+            "checkpoint_id": "2-inner",
+            "checkpoint_ns": "inner",
+        }
+    }
+
+    chkpnt_1: Checkpoint = empty_checkpoint()
+    chkpnt_2: Checkpoint = create_checkpoint(chkpnt_1, {}, 1)
+    chkpnt_3: Checkpoint = empty_checkpoint()
+
+    metadata_1: CheckpointMetadata = {
+        "source": "input",
+        "step": 2,
+        "writes": {},
+        "score": 1,
+    }
+    metadata_2: CheckpointMetadata = {
+        "source": "loop",
+        "step": 1,
+        "writes": {"foo": "bar"},
+        "score": None,
+    }
+    metadata_3: CheckpointMetadata = {}
+
+    return {
+        "configs": [config_1, config_2, config_3],
+        "checkpoints": [chkpnt_1, chkpnt_2, chkpnt_3],
+        "metadata": [metadata_1, metadata_2, metadata_3],
+    }
+
+
+@pytest.mark.parametrize("saver_name", ["base", "pool", "pipe"])
+def test_search(saver_name: str, test_data) -> None:
+    with _saver(saver_name) as saver:
+        configs = test_data["configs"]
+        checkpoints = test_data["checkpoints"]
+        metadata = test_data["metadata"]
+
+        saver.put(configs[0], checkpoints[0], metadata[0], {})
+        saver.put(configs[1], checkpoints[1], metadata[1], {})
+        saver.put(configs[2], checkpoints[2], metadata[2], {})
+
+        # call method / assertions
+        query_1 = {"source": "input"}  # search by 1 key
+        query_2 = {
            "step": 1,
            "writes": {"foo": "bar"},
-            "score": None,
-        }
-        self.metadata_3: CheckpointMetadata = {}
-        with PostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            saver.setup()
+        }  # search by multiple keys
+        query_3: dict[str, Any] = {}  # search by no keys, return all checkpoints
+        query_4 = {"source": "update", "step": 1}  # no match

-    def test_search(self) -> None:
-        with PostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            # save checkpoints
-            saver.put(self.config_1, self.chkpnt_1, self.metadata_1, {})
-            saver.put(self.config_2, self.chkpnt_2, self.metadata_2, {})
-            saver.put(self.config_3, self.chkpnt_3, self.metadata_3, {})
+        search_results_1 = list(saver.list(None, filter=query_1))
+        assert len(search_results_1) == 1
+        assert search_results_1[0].metadata == metadata[0]

-            # call method / assertions
-            query_1 = {"source": "input"}  # search by 1 key
-            query_2 = {
-                "step": 1,
-                "writes": {"foo": "bar"},
-            }  # search by multiple keys
-            query_3: dict[str, Any] = {}  # search by no keys, return all checkpoints
-            query_4 = {"source": "update", "step": 1}  # no match
+        search_results_2 = list(saver.list(None, filter=query_2))
+        assert len(search_results_2) == 1
+        assert search_results_2[0].metadata == metadata[1]

-            search_results_1 = list(saver.list(None, filter=query_1))
-            assert len(search_results_1) == 1
-            assert search_results_1[0].metadata == self.metadata_1
+        search_results_3 = list(saver.list(None, filter=query_3))
+        assert len(search_results_3) == 3

-            search_results_2 = list(saver.list(None, filter=query_2))
-            assert len(search_results_2) == 1
-            assert search_results_2[0].metadata == self.metadata_2
+        search_results_4 = list(saver.list(None, filter=query_4))
+        assert len(search_results_4) == 0

-            search_results_3 = list(saver.list(None, filter=query_3))
-            assert len(search_results_3) == 3
+        # search by config (defaults to checkpoints across all namespaces)
+        search_results_5 = list(saver.list({"configurable": {"thread_id": "thread-2"}}))
+        assert len(search_results_5) == 2
+        assert {
+            search_results_5[0].config["configurable"]["checkpoint_ns"],
+            search_results_5[1].config["configurable"]["checkpoint_ns"],
+        } == {"", "inner"}

-            search_results_4 = list(saver.list(None, filter=query_4))
-            assert len(search_results_4) == 0

-            # search by config (defaults to checkpoints across all namespaces)
-            search_results_5 = list(
-                saver.list({"configurable": {"thread_id": "thread-2"}})
-            )
-            assert len(search_results_5) == 2
-            assert {
-                search_results_5[0].config["configurable"]["checkpoint_ns"],
-                search_results_5[1].config["configurable"]["checkpoint_ns"],
-            } == {"", "inner"}
-
-            # TODO: test before and limit params
-
-    def test_null_chars(self) -> None:
-        with PostgresSaver.from_conn_string(DEFAULT_URI) as saver:
-            config = saver.put(self.config_1, self.chkpnt_1, {"my_key": "\x00abc"}, {})
-            assert saver.get_tuple(config).metadata["my_key"] == "abc"  # type: ignore
-            assert (
-                list(saver.list(None, filter={"my_key": "abc"}))[0].metadata["my_key"]  # type: ignore
-                == "abc"
-            )
+@pytest.mark.parametrize("saver_name", ["base", "pool", "pipe"])
+def test_null_chars(saver_name: str, test_data) -> None:
+    with _saver(saver_name) as saver:
+        config = saver.put(
+            test_data["configs"][0],
+            test_data["checkpoints"][0],
+            {"my_key": "\x00abc"},
+            {},
+        )
+        assert saver.get_tuple(config).metadata["my_key"] == "abc"  # type: ignore
+        assert (
+            list(saver.list(None, filter={"my_key": "abc"}))[0].metadata["my_key"]
+            == "abc"
+        )
@@ -110,7 +110,7 @@ class SqliteSaver(BaseCheckpointSaver[str]):
                check_same_thread=False,
            )
        ) as conn:
-            yield SqliteSaver(conn)
+            yield cls(conn)

    def setup(self) -> None:
        """Set up the checkpoint database.
@@ -137,7 +137,7 @@ class AsyncSqliteSaver(BaseCheckpointSaver[str]):
            AsyncSqliteSaver: A new AsyncSqliteSaver instance.
        """
        async with aiosqlite.connect(conn_string) as conn:
-            yield AsyncSqliteSaver(conn)
+            yield cls(conn)

    def get_tuple(self, config: RunnableConfig) -> Optional[CheckpointTuple]:
        """Get a checkpoint tuple from the database.
@@ -4,11 +4,13 @@
 # TESTING AND COVERAGE
 ######################

+TEST ?= .
+
 test:
-	poetry run pytest tests
+	poetry run pytest $(TEST)

 test_watch:
-	poetry run ptw .
+	poetry run ptw $(TEST)

 ######################
 # LINTING AND FORMATTING
@@ -1,4 +1,3 @@
-import asyncio
 import logging
 import os
 import pickle
@@ -6,7 +5,6 @@ import random
 import shutil
 from collections import defaultdict
 from contextlib import AbstractAsyncContextManager, AbstractContextManager, ExitStack
-from functools import partial
 from types import TracebackType
 from typing import Any, AsyncIterator, Dict, Iterator, Optional, Sequence, Tuple, Type

@@ -372,7 +370,7 @@ class MemorySaver(
            RunnableConfig: The updated config containing the saved writes' timestamp.
        """
        thread_id = config["configurable"]["thread_id"]
-        checkpoint_ns = config["configurable"]["checkpoint_ns"]
+        checkpoint_ns = config["configurable"].get("checkpoint_ns", "")
        checkpoint_id = config["configurable"]["checkpoint_id"]
        outer_key = (thread_id, checkpoint_ns, checkpoint_id)
        outer_writes_ = self.writes.get(outer_key)
@@ -395,9 +393,7 @@ class MemorySaver(
        Returns:
            Optional[CheckpointTuple]: The retrieved checkpoint tuple, or None if no matching checkpoint was found.
        """
-        return await asyncio.get_running_loop().run_in_executor(
-            None, self.get_tuple, config
-        )
+        return self.get_tuple(config)

    async def alist(
        self,
@@ -418,24 +414,8 @@ class MemorySaver(
        Yields:
            AsyncIterator[CheckpointTuple]: An asynchronous iterator of checkpoint tuples.
        """
-        loop = asyncio.get_running_loop()
-        iter = await loop.run_in_executor(
-            None,
-            partial(
-                self.list,
-                before=before,
-                limit=limit,
-                filter=filter,
-            ),
-            config,
-        )
-        while True:
-            # handling StopIteration exception inside coroutine won't work
-            # as expected, so using next() with default value to break the loop
-            if item := await loop.run_in_executor(None, next, iter, None):
-                yield item
-            else:
-                break
+        for item in self.list(config, filter=filter, before=before, limit=limit):
+            yield item

    async def aput(
        self,
@@ -455,9 +435,7 @@ class MemorySaver(
        Returns:
            RunnableConfig: The updated config containing the saved checkpoint's timestamp.
        """
-        return await asyncio.get_running_loop().run_in_executor(
-            None, self.put, config, checkpoint, metadata, new_versions
-        )
+        return self.put(config, checkpoint, metadata, new_versions)

    async def aput_writes(
        self,
@@ -474,10 +452,9 @@ class MemorySaver(
            config (RunnableConfig): The config to associate with the writes.
            writes (List[Tuple[str, Any]]): The writes to save, each as a (channel, value) pair.
            task_id (str): Identifier for the task creating the writes.
+        return self.put_writes(config, writes, task_id)
        """
-        return await asyncio.get_running_loop().run_in_executor(
-            None, self.put_writes, config, writes, task_id
-        )
+        return self.put_writes(config, writes, task_id)

    def get_next_version(self, current: Optional[str], channel: ChannelProtocol) -> str:
        if current is None:
--- a/Show More
+++ b/Show More