🚀

Quick Start

Prerequisites

Before running GraphRAG, make sure you have: - Python 3.10+ - Node.js 18+ - Docker (optional, for containerized deployment) - A TigerGraph Cloud account (free tier available) - At least one LLM API key (Claude, OpenAI, etc.)

Installation

# Clone the repository
git clone https://github.com/MUTHUKUMARAN-K-1/graphrag-inference-hackathon
cd graphrag-inference-hackathon

# Backend setup
pip install -r requirements.txt

# Frontend setup
cd web
npm install
npm run dev

# Or use Docker
docker build -t graphrag .
docker run -p 3000:3000 -p 8000:8000 graphrag

Environment Variables

# web/.env  (copy from web/.env.example)

# TigerGraph
TG_HOST=https://your-instance.i.tgcloud.io
TG_TOKEN=your_bearer_token
TG_GRAPH=GraphRAG

# LLM — set at least one
OPENAI_API_KEY=sk-...
OPENAI_BASE_URL=https://models.botlearn.ai/v1   # optional: botlearn.ai / other proxy
LLM_MODEL=gemini-2.5-flash                      # optional: override default model

# Embeddings (for live TigerGraph retrieval)
HF_TOKEN=hf_...

# Optional additional providers
ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=AI...
GROQ_API_KEY=gsk_...