🧠 GraphMind

Transform your documents into interactive, queryable, and visually engaging knowledge graphs.

Overview

GraphMind is a mono-repo that combines the power of FastAPI, LangGraph, and modern frontend technologies to enable seamless document ingestion, processing, and visualization. With GraphMind, you can:

• Ingest and process documents (PDF, TXT, MD, and more).
• Extract entities, relationships, and metrics using LLMs.
• Store and query data with ChromaDB.
• Visualize knowledge graphs in an intuitive React-based UI.

Repository Structure

GraphMind/
├── backend/       # Backend service built with FastAPI
│   ├── src/       # Core backend logic
│   │   ├── api/               # FastAPI routes and Pydantic models
│   │   ├── components/        # Modular components for ingestion, processing, and graph building
│   │   ├── workflows/         # LangGraph workflows and state management
│   │   ├── config/            # Configuration and logging
│   ├── tests/     # Pytest test suite
│   └── main.py    # FastAPI app entrypoint
├── frontend/      # Frontend application built with Vite + React + TypeScript
│   ├── src/       # React components, hooks, and pages
│   ├── vite.config.ts
│   └── package.json

Key Features

Backend

• Document Ingestion: Supports multiple formats (PDF, TXT, MD, etc.).
• Smart Chunking: Recursive and token-based strategies for text segmentation.
• Semantic Search: Powered by ChromaDB for efficient retrieval.
• Knowledge Graph Extraction: Entity and relationship extraction using LLMs.
• Orchestrated Pipelines: Modular workflows with LangGraph.
• RESTful API: Clean and scalable API endpoints.

Frontend

• Interactive UI: Upload documents, run queries, and visualize knowledge graphs.
• Modern Stack: Built with Vite, React, TypeScript, and TailwindCSS.
• Real-Time Updates: Seamless integration with backend APIs.

Quickstart Guide

Prerequisites

• Backend: Python 3.11, Redis, Docker (optional).
• Frontend: Node.js (v16+), npm or yarn.

Running Locally

Backend

1. Navigate to the backend/ directory.
2. Create and activate a virtual environment:

   python -m venv .venv
   source .venv/bin/activate

3. Create a .env file (refer to backend/README.md for required variables).
4. Install dependencies in editable mode:

   pip install -e .

5. Start the FastAPI server:

   uvicorn main:app --reload

Frontend

1. Navigate to the frontend/ directory.
2. Create a .env file with the following content:

   VITE_API_BASE_URL=http://localhost:8000/api/v1

3. Install dependencies:

   npm install

4. Start the development server:

   npm run dev

Access the Application

• Backend: API documentation available at http://localhost:8000/docs.
• Frontend: Open http://localhost:5173 in your browser.

Deployment

GraphMind is designed for deployment on modern platforms:

• Frontend: Deploy on Vercel for seamless CI/CD.
• Backend: Deploy on Render with Redis for task management.

Refer to the docs/DEPLOYMENT.md for detailed instructions.

Additional Resources

• Backend Documentation: backend/README.md
• Frontend Documentation: frontend/README.md
• Redis Optimization: docs/REDIS_OPTIMIZATION.md

Notes

• LLM-powered features require an API key. Refer to backend/README.md for environment variable setup.
• ChromaDB supports both embedded (persistent client) and remote configurations.

---

Crafted with ❤️ by the GraphMind team.