Convert TXT files and URLs into interactive knowledge graphs with AI-powered Q&A.
Tech Stack: React + FastAPI + Neo4j + LangChain
Prerequisites: Docker, Docker Compose, 4GB+ RAM
git clone git@github.com:guptaachin/doc2graph.git
cd doc2graph
docker-compose up --buildAccess:
- Frontend: http://localhost:3000
- API: http://localhost:8000
- Neo4j: http://localhost:7474 (neo4j/neo4j1234)
- 📄 Upload TXT files or URLs
- 🗺️ Interactive graph visualization
- ❓ Natural language Q&A
- 🔍 Graph exploration
- Ingest → Upload files/URLs
- Graph → Explore connections
- Q&A → Ask questions
# Database only
docker-compose up neo4j -d
# Backend
cd backend && pip install -r requirements.txt && uvicorn main:app --reload
# Frontend
cd frontend && npm install && npm startCommon fixes:
- Port conflicts: Check 3000, 8000, 7474, 7687
- Memory issues: Close other apps
- Build failures:
docker system prune
Reset:
docker-compose down -v
docker-compose up --buildLogs:
docker-compose logs -f