automatic-tool-generation

Guide: How to Use as a Package

1. Copy into your project
   Clone or copy the automatic-tool-generation module into your project directory.

2. Initialize and run the agent
   Use the following code to run the agent from within your Python script:

   import uuid
   from langgraph.checkpoint.memory import MemorySaver
   from core.graph import builder
   import asyncio
   from dotenv import load_dotenv
   
   load_dotenv()
   
   file_path = "<your_file_path_here>"
   
   agent = AgentBuilder()
   results = asyncio.run(agent.run(file_path))
   

3. (Optional) Set custom configuration
   You can customize how the agent behaves by passing arguments to AgentBuilder:

   class AgentBuilder:
       def __init__(self,
                    planner_provider="openai",
                    planner_model="o3-mini",
                    writer_provider="openai",
                    writer_model="gpt-4.1-nano",
                    file_type="CSV",
                    index4file="true",
                    breadth=2):
           ...
   

   Adjust parameters like planner_model, writer_model, or breadth to fit your use case.

This site is open source. Improve this page.