SelfClaw

SelfClaw is a sophisticated multi-model AI aggregation system designed for seamless integration and enhanced performance in dialogue applications. Its key features include:

• Multi-Model Aggregation: Support for various AI models such as DeepSeek, NVIDIA, SiliconFlow, ZenMux, Zhipu, Kimi, MiMo, and Mistral allows for diverse functionalities tailored to user needs.
• Automatic Fallback: The system intelligently manages model failures by automatically switching to alternative models, retrying temporary failures, providing a 24-hour blacklist for persistent issues, and imposing a cooldown period during system-wide downtimes.
• CCR Closed-Loop Agent: Incorporates advanced capabilities including tool invocation, memory management, reflective repair processes, and strategy evolution to optimize responses.
• File Management: Offers features to read, write, and list directories, with support for parsing PDFs, Word documents, and Excel files.
• Internet Search: Facilitates dual-engine searches across Bing and Baidu to enhance information retrieval.
• Voice Interaction: Integrates voice output through edge-tts and allows for speech input with the speech_recognition library.
• Model Management: Provides a user-friendly interface for visualizing and managing models, including options for adding, deleting, and reordering them.
• Custom Themes: Users can switch between dark, soft gray, and light themes to enhance their experience.
• CCR HTTP Proxy: Compatible with Claude Code request formats, enabling automatic forwarding to the OpenClaw gateway.
• Self-IDE: Features a local AI Agent IDE service with a VS Code extension, allowing for streamlined development and testing processes.

Project Structure

The repository includes a well-organized directory structure featuring modules for UI operations, a closed-loop agent, and more:

MyAI_Model/
├── DS/                    # Streamlit UI Module
│   ├── main.py            # Entry point
│   ├── config.py          # Configuration management
│   ├── chat_engine.py     # Dialogue engine (single-use/auto-fallback)
│   ├── model_manager.py   # Model CRUD operations, blacklisting, logging
│   ├── tools.py           # File operations, search, memory management, plugins
│   └── ui.py              # Complete UI rendering (CSS + components)
├── CCR/                   # CCR closed-loop agent
│   └── ccr.py             # Hermes agent + HTTP proxy
├── gateway.py             # FastAPI gateway (multi-model routing)
├── openclaw.json          # Model pool & Provider configurations
├── start.bat              # One-click launch script for Windows
├── requirements.txt       # Python dependencies
└── README.md

Technology Stack

The project employs modern technologies including Streamlit for the front-end, FastAPI and Uvicorn for the backend gateway, and OpenAI SDK for AI functionalities. Advanced voice processing and dual-engine search capabilities further enhance the overall experience.

SelfClaw represents a comprehensive and flexible solution for developers and organizations looking to implement advanced AI aggregation systems in their applications.