Automatic Tool Generation

Note

This project is designed to work seamlessly with create_react_agent from LangGraph. The generated tool functions are fully compatible with the create_react_agent framework. However, if you remove the @tool decorator, these functions can also be used in any other framework.

This is a subproject of Create Data Scientist Agent

The main goal of this project is to automatically and independently generate tools for use by LLM agents. Rather than manually predefining tools for every possible use case, this project introduces a pipeline that can:

1. Analyze the task type
2. Ideate potential solutions
3. Generate appropriate Python tools
4. Validate them against sample data

This approach enables more flexible, scalable, and intelligent agents that can adapt to diverse data analysis scenarios without requiring human experts in tool preparation. There are some examples of generated tools avaliable.

Warning

CLI tools are still under active development, and functionality is currently limited. For full experience, consider using LangGraph Studio instead.

👉 Check how to get started with LangGraph Studio for this project

Quick Start

1. Set up your environment

   cp .env.example .env

   Update the .env file with your own API keys and configuration as needed.

2. Install dependencies

   pip install -r requirements.txt

3. (Optional) Use LangGraph Studio
   If you want to use LangGraph Studio for a better development experience:

   pip install 'langgraph-cli[inmem]'

   This step does not require creating account. It run a local server and open a web browser to control the agent

View the documntation for more details on features and configuration options best suited to your project.

Table of Content

• Usage Guideline:
  • LangGraph Dev(recommand)
  • Cli tool
  • Use in Your Project