Repolode
Submit
langchain-ai

langchain-ai/langgraph-bigtool

A Python library for building LangGraph agents that search a store to retrieve and use hundreds or thousands of tools.

View on GitHub
Stars
559
Forks
49
Open issues
13
Last commit
2 days ago

What is it?

langgraph-bigtool is a Python library for creating LangGraph agents that can access large numbers of tools. Instead of passing every tool to the language model at once, it stores tool descriptions in LangGraph's long-term memory store and gives the agent a retrieval tool that searches for the tools relevant to the current problem.

Tool metadata can be kept in LangGraph's persistence layer, with in-memory and Postgres backends available. Retrieval is customizable: you can pass your own function to find tools, such as one that runs semantic search, or one that selects tools by category without using a store at all.

The library is built on LangGraph and supports its streaming, short-term and long-term memory, and human-in-the-loop features. It is installed with pip and includes a quickstart that equips an agent with the functions from Python's math module.

Key features

  • 01Gives agents access to hundreds or thousands of tools through retrieval rather than passing all tools at once.
  • 02Stores tool names, descriptions and namespaces in LangGraph's persistence layer.
  • 03Supports in-memory and Postgres storage backends.
  • 04Lets you replace the default retrieval with custom search or category-based logic.
  • 05Works with LangGraph streaming, memory and human-in-the-loop features.

Use cases

  • Building an agent that must choose from a large catalog of API or internal tools.
  • Letting an LLM work with tool sets too large to fit in a single prompt.
  • Organizing tools into categories so the model selects a group before a specific tool.

Quick start

  1. 1

    Install the library from PyPI

    pip install langgraph-bigtool
  2. 2

    Install the OpenAI extras used by the quickstart example

    pip install langgraph-bigtool "langchain[openai]"
  3. 3

    Set your OpenAI API key in the environment

    export OPENAI_API_KEY=<your_api_key>
  4. 4

    Run the quickstart Python example that equips an agent with math library tools

Things to know

  • The quickstart example requires an OpenAI API key and calls paid OpenAI models.
  • Semantic tool retrieval depends on an embedding model and a store, so setup involves more than the core library.

Topics

#langgraph#llm-agents#tool-calling#retrieval#semantic-search#python

FAQ

What is langgraph-bigtool?
A Python library for building LangGraph agents that search a store to retrieve and use hundreds or thousands of tools. langgraph-bigtool is a Python library for creating LangGraph agents that can access large numbers of tools. Instead of passing every tool to the language model at once, it stores tool descriptions in LangGraph's long-term memory store and gives the agent a retrieval tool that searches for the tools relevant to the current problem.
Is langgraph-bigtool free to use? What license does it have?
langgraph-bigtool is open source under the MIT license. Check the license terms before using it in commercial products.
Is langgraph-bigtool still maintained?
Actively maintained. The last commit to langgraph-bigtool was on Nov 5, 2026.
What can I use langgraph-bigtool for?
Building an agent that must choose from a large catalog of API or internal tools. Letting an LLM work with tool sets too large to fit in a single prompt. Organizing tools into categories so the model selects a group before a specific tool.
How do I install langgraph-bigtool?
Install the library from PyPI: pip install langgraph-bigtool Install the OpenAI extras used by the quickstart example: pip install langgraph-bigtool "langchain[openai]" Set your OpenAI API key in the environment: export OPENAI_API_KEY=<your_api_key> Run the quickstart Python example that equips an agent with math library tools
Where can I get langgraph-bigtool?
The source code and installation instructions are on GitHub: https://github.com/langchain-ai/langgraph-bigtool

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects