This project provides an open-source ecosystem for ChatGPT, enabling it to interact with various tools and execute commands via natural language. It aims to expand ChatGPT's capabilities beyond text generation, allowing users to automate tasks, access real-time information, and control their computer, thereby boosting productivity. The target audience includes developers and power users looking to integrate LLMs with external functionalities.

How It Works

The core mechanism is based on Chain-of-Thought (CoT) prompting. The system provides ChatGPT with a list of available tools and their descriptions. ChatGPT, guided by specific prompts and a defined communication protocol (Thought, Action, Action Input), determines which tool to use and its parameters. The engine parses ChatGPT's output, executes the corresponding Python functions (which wrap libraries like subprocess, requests, or the Python interpreter), and feeds the results back to ChatGPT for iterative reasoning or final output.

Quick Start & Requirements

• Install dependencies: pip3 install -r requirements.txt
• Configure API keys and settings in .env and config.json.
• Run the main script: python3 terminal_io.py
• Requires Python 3.x and an OpenAI API key.
• Official demo: LLM-OS demo
• Tool plugin tutorial: tool插件使用教程

Highlighted Details

• Supports natural language interaction in both Chinese and English.
• Features context memory for conversational continuity.
• Enables automatic, sequential, and tree-like orchestration of multiple tools.
• Offers proxy support for network requests.
• Can be integrated into other projects via PyPI (pip install chatgpt-tool-hub).

Maintenance & Community

• The project is inspired by and has rewritten parts of Langchain and Auto-GPT.
• Community discussion and suggestions are encouraged via GitHub Issues.
• Support the author via donations (☕).

Licensing & Compatibility

• The README does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

Executing arbitrary code via tools like terminal and python grants ChatGPT significant control over the local machine, posing potential security risks. Users who do not trust ChatGPT's execution should avoid using these tools. The system's robustness against low-quality prompts or inappropriate tool inputs is handled, but not guaranteed to be 100%.