llm-hub/README.md

61 lines
1.5 KiB
Markdown

# 🤖 Agentic LLM Hub
Self-hosted AI agent platform with multi-provider LLM aggregation, reasoning engines (ReAct, Plan-and-Execute, Reflexion), MCP tools, and web IDE.
## 🚀 Quick Start
```bash
# 1. Clone from your Gitea
git clone https://gitea.yourdomain.com/youruser/llm-hub.git
cd llm-hub
# 2. Configure
cp .env.example .env
nano .env # Add your API keys
# 3. Deploy
./setup.sh && ./start.sh
```
## 📡 Access Points
| Service | URL | Description |
|---------|-----|-------------|
| VS Code IDE | `http://your-ip:8443` | Full IDE with Continue.dev |
| Agent API | `http://your-ip:8080/v1` | Main API endpoint |
| LiteLLM | `http://your-ip:4000` | LLM Gateway |
| MCP Tools | `http://your-ip:8001/docs` | Tool OpenAPI docs |
| ChromaDB | `http://your-ip:8000` | Vector memory |
| Web UI | `http://your-ip:3000` | Chat interface |
## 🔧 Supported Providers
- **Groq** (Free tier, fast)
- **Mistral** (1B tokens/month free)
- **Anthropic Claude** (Trial credits)
- **Moonshot Kimi** ($5 signup bonus)
- **OpenRouter** (Free tier access)
- **Cohere** (1K calls/month)
- **DeepSeek** (Cheap reasoning)
## 🧠 Reasoning Modes
- `react` - Fast iterative reasoning
- `plan_execute` - Complex multi-step tasks
- `reflexion` - Self-correcting with verification
- `auto` - Automatic selection
## 📚 Documentation
- [Setup Guide](docs/SETUP.md)
- [API Reference](docs/API.md)
- [Provider Guide](docs/PROVIDERS.md)
## 🔄 Updates
```bash
git pull origin main
docker-compose pull
docker-compose up -d
```