Torah Study AI
April 9, 2026 ยท View on GitHub
Your AI study partner for Torah. Real sources. Every answer.
An AI chavruta that helps you explore Torah, Talmud, and Jewish texts with verifiable sources from Sefaria.
Features
- Chat with sources - Ask a question, get an answer with clickable Sefaria links
- Conversation history - Sidebar with past chats, resume any conversation
- Multi-language - French interface, Hebrew quotes, answer in your language
- Parashah of the week - Weekly Torah portion with key insights
Tech Stack
| Layer | Technology |
|---|---|
| Frontend | Next.js + shadcn/ui |
| Backend | FastAPI (Python) |
| Database | SQLite |
| Auth | JWT + bcrypt |
| RAG | LangChain + Weaviate + Cohere Rerank |
| Embeddings | Gemini Embedding 001 |
| LLM | Gemini 2.5 Flash |
| Data | Sefaria HuggingFace datasets (3.5M texts) |
Quick Start
# Clone
git clone https://github.com/davalimi/torah-study-ai.git
cd torah-study-ai
# Set up environment
cp .env.example .env
# Edit .env with your GOOGLE_API_KEY
# Install
make install
make install-frontend
# Run (2 terminals)
make api # Terminal 1: API on localhost:8000
make frontend # Terminal 2: Frontend on localhost:3000
Project Structure
torah-study-ai/
src/
api/ # FastAPI backend (auth, chat, sessions)
frontend/ # Next.js + shadcn/ui
rag/ # LangChain RAG pipeline (Block 3)
scripts/ # Download Sefaria, indexing, evaluation
tests/ # pytest (18 tests)
data/ # SQLite DB + local data (gitignored)
Documentation
Full project documentation on skillia.dev/projects/torah-study-ai
License
MIT