Centauri Chat Service is a lightweight, scalable WebSocket-based AI chat service. It allows multiple conversational sessions to be maintained in memory, with a focus on modularity and extendibility. Built using FastAPI, it follows a Hexagonal Architecture to ensure flexibility and maintainability.

The service uses an in-memory singleton pattern to store data, suitable for low-demand scenarios. However, the architecture is designed to allow easy integration with Redis or traditional databases if higher scalability is required.

• WebSocket Support: Real-time communication through /chat/ and /chat/{chat_id} endpoints.
• Session Management: Multiple conversations stored and managed in memory.
• Health Check Endpoint: /check to verify service status.
• Hexagonal Architecture: Clear separation of concerns, enabling future extensibility.
• In-Memory Storage: Uses a singleton pattern for low-demand use cases.
• Extendable to Redis or Databases: Ready for scaling with minimal changes.

1. Python: Version 3.9 or higher.
2. Docker: Ensure Docker is installed and running.
3. Make: For convenience in executing commands (optional, but recommended).

Follow these steps to get the project up and running:

git clone [email protected]:Endika/Centauri.git
cd Centauri

echo "OPENAI_API_KEY=your-token" > .env

make build
make up
# or
docker-compose up --build -d

make chat
# or
docker exec -it centauri-client wscat -c ws://centauri:8000/chat/

make chat-flight
# or
docker exec -it centauri-client wscat -c ws://centauri:8000/chat/flight_attendant