Multi-model AI assistant platform with web client and API backend.
nexumi-ai/
apps/web/ React frontend with TypeScript and Tailwind
backend/ Express API server with proxy endpoints
git clone https://github.com/rexblade58/nexumi-ai.git
cd nexumi-ai
npm install
npm run dev
- Chat interface with streaming responses
- Multi-model support via OpenRouter and Together AI
- Local model support through Ollama
- Image generation and analysis endpoints
- Offline detection with graceful fallback
- Web worker for background AI processing
| Endpoint |
Method |
Description |
/health |
GET |
Health check and uptime |
/api/chat |
POST |
Chat completion proxy |
/api/generate-image |
POST |
Image generation via FLUX.1 |
/api/analyze-image |
POST |
Image analysis via Gemini |
| Layer |
Technologies |
| Frontend |
React 18, TypeScript, Vite, Tailwind CSS, Zustand |
| Backend |
Node.js, Express, CORS, Axios |
| AI |
OpenAI SDK, OpenRouter, Together AI, Ollama |
| Branch |
Purpose |
main |
Production releases |
develop |
Active development and integration |
staging |
Pre-release testing and QA |
MIT (c) 2021-2025 Menard Rosal
