📄 pdfGPT

AI-powered PDF document Q&A: upload a PDF, ask questions, and get grounded answers pulled from your document via semantic search with vector embeddings and a fast LLM.

✨ Features

• PDF upload and text extraction
• Chunking with overlap for robust retrieval
• Sentence-transformer embeddings (384-d) on Hugging Face Inference API
• Vector similarity search in Supabase (pgvector)
• Context-grounded answers with Groq LLM (Llama 3.3)
• Clean UI built with Next.js App Router and React 19

🧰 Tech Stack

• Frontend: Next.js 15, React 19
• AI/RAG: LangChain (loaders, splitters, embeddings), Hugging Face Inference
• Vector DB: Supabase (Postgres + pgvector)
• LLM: Groq SDK (Llama-3.3-70B Versatile)

---

🚀 Getting Started

1) Clone and install

git clone <your-repo-url>
cd pdfgpt
npm install

2) Environment variables

Create a .env.local at the project root with:

# Hugging Face Inference (free tier works)
HUGGINGFACE_API_KEY=your_hf_api_key

# Supabase (Project Settings → API)
SUPABASE_URL=your_supabase_project_url
SUPABASE_PRIVATE_KEY=your_service_role_key

# Groq (https://console.groq.com/)
GROQ_API_KEY=your_groq_api_key