Skip to content

README.md

Latest commit

 

History

History
33 lines (23 loc) · 1.67 KB

README.md

File metadata and controls

33 lines (23 loc) · 1.67 KB

Loop AI

Monorepo for the Loop AI project. Frontend lives in frontend/, backend in backend/.

Tech stack

  • Frontend: React, TypeScript, Vite, Tailwind CSS, Radix UI, Zustand, React Router, TanStack Query, Supabase (auth).
  • Backend: FastAPI, Uvicorn, Supabase (Postgres + Auth), Redis, RQ (background jobs).
  • LLM: OpenRouter (streaming chat completions).

Backend (FastAPI + Supabase + Redis)

  • Supabase setup: See docs/SUPABASE_SETUP.md for full setup (CLI, migrations, bucket, env).
  • Run API (from backend/): uvicorn app.main:app --reload --port 4000
  • Run worker (from backend/): python worker.py
  • Requires Redis running and a root .env (copy from backend/.env.example or root .env.example). The root .env must include:
    • Supabase: SUPABASE_URL, SUPABASE_ANON_KEY, SUPABASE_SERVICE_ROLE_KEY
    • REDIS_URL (e.g. redis://localhost:6379)
    • OpenRouter: OPENROUTER_API_KEY (get one at openrouter.ai/keys) so the LLM works. Optional: OPENROUTER_MODEL, OPENROUTER_MAX_TOKENS, OPENROUTER_TEMPERATURE (see backend/.env.example).

Frontend

cd frontend
npm install
npm run dev

App runs at http://localhost:8080.

Set up frontend/.env from frontend/.env.example: VITE_SUPABASE_URL, VITE_SUPABASE_ANON_KEY, and VITE_API_URL=http://localhost:4000 (and VITE_BACKEND_WS_URL=ws://localhost:4000/ws if using streaming). You can sync from the root .env with node scripts/sync-env.js (see docs/GETTING_STARTED.md).

Other scripts (from frontend/): npm run build, npm run preview, npm run lint, npm run test.