Turn any content into an interactive AI teacher. Learning that talks back.
| Layer | Tool | Free tier |
|---|---|---|
| Frontend | Next.js 14 on Vercel | Yes |
| Backend | FastAPI on Render | Yes |
| Database + Auth | Supabase | Yes |
| Vector store | Supabase pgvector | Yes (built-in) |
| File storage | Supabase Storage | Yes |
| LLM | Claude API (Anthropic) | Trial credits |
learnflow/
├── frontend/ # Next.js app (deploy to Vercel)
│ ├── app/
│ │ ├── page.tsx # Landing page
│ │ ├── upload/page.tsx # Upload content
│ │ └── teacher/page.tsx # Chat with AI teacher
│ ├── components/
│ │ ├── chat/ # Chat UI components
│ │ ├── upload/ # Upload UI components
│ │ └── ui/ # Shared UI components
│ └── lib/
│ ├── supabase.ts # Supabase client
│ └── api.ts # Backend API calls
│
└── backend/ # FastAPI app (deploy to Render)
├── main.py # App entry point
├── routers/
│ ├── ingest.py # POST /ingest — upload + parse content
│ └── chat.py # POST /chat — AI teacher chat
├── services/
│ ├── parser.py # PDF, video, text parsers
│ ├── chunker.py # Text chunking logic
│ ├── embedder.py # Embed chunks → pgvector
│ └── teacher.py # RAG + Claude chat logic
├── models/
│ └── schemas.py # Pydantic request/response models
└── utils/
└── supabase.py # Supabase client (backend)
git clone <your-repo>
cd learnflow- Go to https://supabase.com and create a free project
- Run the SQL in
backend/utils/schema.sqlin the Supabase SQL editor - Enable pgvector: in SQL editor run
create extension vector; - Copy your project URL and anon key
cd backend
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env # fill in your keys
uvicorn main:app --reloadcd frontend
npm install
cp .env.example .env.local # fill in your keys
npm run dev- Frontend → push to GitHub, import on Vercel, add env vars
- Backend → push to GitHub, create Web Service on Render, add env vars