This is the backend service for m3-chat, a 100% open-source and free AI chat web app powered by local AI models using Ollama.
The backend exposes an API endpoint that proxies requests to local AI models, streams responses back to the frontend in real-time, and manages communication with the Ollama CLI.
- Tech stack: TypeScript, Zod, ElysiaJS, ngrok
- Streaming: Uses child process streams for live AI responses
- Port: 2000 by default
- CORS enabled for frontend integration
Query Parameters:
model(string): The AI model to use (e.g.,llama3:8b)content(string): The prompt or message content
Streams the AI-generated text response as plain text with chunked transfer encoding.
- Use infra