AgentKit

Modern multi-agent AI assistant with LangGraph & AG-UI

🌐 Tiếng Việt | English

Build powerful AI applications with real-time streaming, interactive UI components, and multi-agent orchestration.

✨ Features

• 🎛️ A2UI Protocol: Agents generate interactive UI components (checkboxes, forms, buttons) in chat
• 🎨 Canvas Mode: Full-screen workspace for complex agent workflows
• 🔄 Multi-Agent Orchestration: LangGraph-powered workflow management
• ⚡ Real-time Streaming: AG-UI protocol with instant agent feedback
• 💬 Thread Management: Persistent conversation threads with SQLite/PostgreSQL
• 🎯 Modern Stack: NextJS + Shadcn UI + FastAPI + LangGraph
• 🔌 Multi-LLM Support: Ollama, Azure OpenAI, Gemini, and more
• 🛡️ Type-Safe: TypeScript frontend, Python type hints backend
• 📊 Observability: Optional LangFuse integration

🚀 Getting Started

Prerequisites

1. Python 3.10+ and Node.js 18+
2. Ollama (or configure Azure OpenAI/Gemini):

   curl -fsSL https://ollama.com/install.sh | sh
   ollama pull qwen:7b
   

Backend Setup

cd backend

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Configure environment (optional)
cp .env.example .env
# Edit .env for Azure OpenAI, Gemini, or PostgreSQL

# Run database migrations
python migrate.py

# Start server
python main.py

✅ Backend running at http://localhost:8000

Frontend Setup

cd frontend

# Install dependencies
npm install

# Start development server
npm run dev

✅ Frontend running at http://localhost:3000

🗄️ Database Configuration

SQLite (default - zero config):

DATABASE_URL=sqlite+aiosqlite:///./agentkit.db

PostgreSQL (production):

DATABASE_URL=postgresql+asyncpg://user:password@localhost:5432/agentkit

Run migrations after any config change:

python migrate.py

🔌 LLM Provider Configuration

Edit .env in backend directory:

Ollama (default):

DEFAULT_PROVIDER=ollama
OLLAMA_BASE_URL=http://localhost:11434
OLLAMA_MODEL=qwen:7b

Azure OpenAI:

DEFAULT_PROVIDER=azure-openai
AZURE_OPENAI_API_KEY=your_key
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com
AZURE_OPENAI_DEPLOYMENT=gpt-5-mini

Gemini:

DEFAULT_PROVIDER=gemini
GEMINI_API_KEY=your_key
GEMINI_MODEL=gemini-2.5-flash

🏗️ Architecture

┌─────────────┐         ┌──────────────┐         ┌─────────────┐
│   Frontend  │◄────────│   AG-UI      │────────►│   Backend   │
│  (NextJS)   │  SSE    │   Protocol   │  HTTP   │  (FastAPI)  │
└─────────────┘         └──────────────┘         └─────────────┘
                                                          │
                                                          ▼
                                                   ┌─────────────┐
                                                   │  LangGraph  │
                                                   │   Agents    │
                                                   └─────────────┘
                                                          │
                                                          ▼
                                                   ┌─────────────┐
                                                   │  LLM Provider│
                                                   │  (Ollama/   │
                                                   │   Azure/    │
                                                   │   Gemini)   │
                                                   └─────────────┘

Backend Stack:

• FastAPI + LangGraph + AG-UI Protocol
• SQLAlchemy (async) + SQLite/PostgreSQL
• Ollama/Azure OpenAI/Gemini

Frontend Stack:

• NextJS 14 + TypeScript + Shadcn UI
• AG-UI client for event streams
• LocalStorage for thread persistence

📚 Documentation

• Multi-Agent Architecture
• Database Setup
• A2UI Protocol
• Canvas Mode
• Implementation Plans
• Knowledge Base

🛠️ Development

Project Structure

agentkit/
├── backend/              # FastAPI + LangGraph backend
│   ├── agents/          # Agent implementations
│   ├── graphs/          # LangGraph workflows
│   ├── api/             # REST endpoints
│   ├── database/        # SQLAlchemy models & migrations
│   ├── llm/             # LLM provider integrations
│   ├── protocols/       # AG-UI protocol implementation
│   └── main.py          # Entry point
├── frontend/            # NextJS frontend
│   ├── app/            # App router pages
│   ├── components/     # React components
│   ├── services/       # API client
│   └── types/          # TypeScript types
└── .docs/              # Documentation

API Endpoints

Chat: POST /api/chat - Stream agent responses Threads: GET/POST /api/threads - Manage conversations Messages: GET /api/threads/{id}/messages - Get thread history Health: GET /health - Server status

Adding New Agents

1. Create agent in backend/agents/
2. Define graph in backend/graphs/
3. Register in backend/agents/agent_registry.py
4. Update frontend components if needed

See agents.md for detailed guide.

🧪 Testing

Backend:

cd backend
pytest

Frontend:

cd frontend
npm test

📝 License

MIT License - see LICENSE file for details

🤝 Contributing

Contributions welcome! Please read our contributing guidelines first.

🌟 Show Your Support

If you find AgentKit useful, please consider giving it a star ⭐️