LyNexus - Community-Driven AI Agent Platform

Empower your AI development with community-driven intelligence

简体中文

What is LyNexus?

LyNexus is a community-driven AI agent platform that empowers developers to create, customize, and share intelligent agents with configurable prompts, tools, and behavior parameters. Built with flexibility and collaboration in mind, it transforms AI development into a shared experience.

Key Philosophy: Create once, share everywhere. Learn from others, build faster.

Features

Flexible AI Configuration

• Model Agnostic: Connect to various AI models (DeepSeek, OpenAI, Anthropic, etc.)
• Parameter Freedom: Fine-tune temperature, tokens, penalties, and more
• Custom System Prompts: Tailor AI behavior to your exact needs
• MCP Integration: Seamlessly integrate Model Context Protocol tools

Modern Interface

• Built with React + Vite + Tailwind CSS
• Telegram-style chat interface
• Real-time streaming responses
• Dark theme optimized for long sessions
• File drag-and-drop support
• Multi-conversation management
• Export chats as TXT, JSON, or Markdown

Community-Driven Ecosystem

• One-Click Export: Package your entire configuration for sharing
• One-Click Import: Use others' optimized setups instantly
• Configuration Sharing: Share & discover configurations
• Stand on Giants' Shoulders: Build upon community-tested configurations

Advanced Tooling

• Dynamic Tool Discovery: AI automatically understands and uses MCP tools through their descriptions - no hardcoded tool names
• Command Execution: AI can execute shell commands (configurable)
• Tool Integration: Connect to various APIs and services via MCP protocol
• Execution Control: Stop long-running operations anytime
• History Management: Preserve and load conversation history

Quick Start

Development Mode

# Terminal 1: Start API server
python -m uvicorn api_server:app --reload

# Terminal 2: Start WebUI
cd webui
npm run dev

Open http://localhost:5173 in your browser.

Release Version

Coming soon - Pre-built release packages will be available for download.

Project Structure

LyNexus/
├── api_server.py           # FastAPI backend server
├── aiclass.py              # Core AI functionality
├── mcp_utils.py            # MCP protocol utilities
│
├── webui/                  # Modern WebUI (React + Vite)
│   ├── src/
│   │   ├── components/     # React components
│   │   ├── services/       # API client & services
│   │   └── stores/         # State management (Zustand)
│   ├── electron/           # Electron desktop wrapper
│   ├── package.json
│   └── vite.config.ts
│
├── tools/                  # MCP Tools Implementation
│   ├── files.py            # File operations
│   ├── network.py          # Network requests
│   ├── ocr.py              # OCR functionality
│   └── osmanager.py        # OS operations
│
├── utils/                  # Utility Functions
│   ├── config_manager.py   # Configuration management
│   ├── chat_data_manager.py # Chat data handling
│   ├── stream_processor.py # Stream processing
│   └── markdown_renderer.py # Markdown rendering
│
├── config/                 # Configuration files
├── data/                   # User data directory
│   └── conversations/      # Conversation storage
├── docs/                   # Documentation
│   ├── README.md           # English documentation
│   └── zh-cn/README.md     # Chinese documentation
│
├── requirements.txt        # Python dependencies
├── pyproject.toml          # Project configuration
└── uv.lock                 # Dependency lock file

Documentation

For detailed installation, configuration, and usage instructions, visit our documentation website.

The documentation covers:

• Installation and setup
• Configuration options
• Usage guide
• Sharing workflows
• Advanced features
• API reference
• And much more!

API Endpoints

Once the API server is running, visit:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Development

Start Backend Only

python -m uvicorn api_server:app --reload

Start Frontend Only

cd webui
npm run dev

Build WebUI

cd webui
npm run build

Build Electron App

cd webui
npm run electron:build

Data Location

All conversations and settings are stored in:

• data/conversations/{id}/ - Per-conversation data
  • settings.json - Conversation settings
  • .confignore - API key (not exported)
  • {id}_ai.json - Message history
  • tools/ - MCP tool files

Community & Contribution

Join Our Community

• GitHub Issues: Report bugs and request features
• GitHub Discussions: Share your amazing AI agents
• Documentation: https://solynacversion.github.io/LyNexus

Contributing

We welcome contributions! Here's how you can help:

1. Share Configurations: Upload your best setups
2. Report Bugs: Help us improve stability
3. Suggest Features: What would make LyNexus better?
4. Improve Documentation: Help others get started
5. Code Contributions: Submit pull requests

Check out our Contributing Guidelines for details.

Troubleshooting

Port 8000 Already in Use

# Windows
netstat -ano | findstr :8000
taskkill /PID <PID> /F

# Linux/Mac
lsof -ti:8000 | xargs kill -9

Port 5173 Already in Use

# Windows
netstat -ano | findstr :5173
taskkill /PID <PID> /F

# Linux/Mac
lsof -ti:5173 | xargs kill -9

Backend Not Connecting

1. Check if API server is running: http://localhost:8000/docs
2. Check WebUI console (F12) for errors
3. Verify webui/.env has correct VITE_API_BASE_URL

"AI Not Initialized" Error

1. Go to Settings (gear icon in sidebar)
2. Enter your API key
3. Click "Save Settings"

License

This project is licensed under the Mozilla Public License 2.0 - see the LICENSE file for details.

Acknowledgments

• AI Community: For sharing knowledge and configurations
• Open Source Projects: That made this platform possible
• Contributors: Everyone who helps improve LyNexus
• Users: For trusting us with your AI development needs

---

Built with passion by the AI community, for the AI community

Documentation • Issues • Discussions

_{Star this repository if you find it useful!}