RAG Document Chat System

A Retrieval-Augmented Generation (RAG) system using Pinecone, Ollama, and Streamlit.

Features

• 📄 Multi-document upload (PDF, TXT, MD, etc.)
• 🔍 Semantic search with vector embeddings
• 💬 Chat with your documents using local LLM
• 🎨 Beautiful Streamlit UI

Setup

1. Install dependencies:

uv sync

2. Set up Pinecone:
   • Sign up at https://pinecone.io
   • Get your API key
   • Create .streamlit/secrets.toml:

   PINECONE_API_KEY = "your_key_here"

3. Install Ollama models:

ollama pull hf.co/CompendiumLabs/bge-base-en-v1.5-gguf
ollama pull hf.co/bartowski/Llama-3.2-1B-Instruct-GGUF

4. Run the app:

streamlit run main.py

last upadted 11 Jan 2026 at 9:47 PM Sunday 9:47 PM

Still in progress..Testing and running right now.

Usage

1. Upload documents via sidebar
2. Click "Process Documents"
3. Start chatting!