Knowledge Graph Assistant (BrainGraph)

An interactive, AI-powered knowledge graph that transforms research PDFs into a visual landscape of authors, papers, and thematic connections.

🚀 Features

• Topic-Centric Visualization: Built using D3.js with a focus on deep thematic research topics.
• LLM-Powered Extraction: Uses Groq (120B model) for lightning-fast, high-reasoning PDF analysis.
• Smart Similarity: Automatically identifies and merges related research concepts across different papers.
• Agentic Workflow: Powered by LangGraph for robust, stateful PDF processing and refinement.
• Multi-Project Support: Save, rename, and load multiple graph projects as JSON files.

🛠️ Tech Stack

• Orchestration: LangGraph
• Backend: FastAPI (Python)
• Frontend: Vanilla JS, D3.js, CSS3
• PDF Processing: PyMuPDF
• LLM: Groq API (openai/gpt-oss-120b)
• Graph Logic: NetworkX

📋 Setup

1. Install Dependencies:

   pip install -r src/backend/requirements.txt

2. Configure API Keys: Create a .env file based on .env.example:

   GROQ_API_KEY=your_key_here

3. Run the Application:

   python src/backend/main.py

   Open http://localhost:8000 in your browser.

📁 Project Structure

• src/backend/: Logic for PDF extraction and graph construction.
• src/frontend/: D3.js visualization and dashboard UI.
• data/docs/: Drop your PDFs here for automatic processing.
• data/: Stores your saved graph projects (.json).

---

Coded with 🤖 Google Antigravity.