Status: [UNTESTED] Endpoints found: http://localhost:8000/api/v1/query Source: https://github.com/swapnilparate87/Enterprise-RAG-System # 🧠 Enterprise RAG System > A 100% FREE, production-ready Retrieval Augmented Generation (RAG) system built with local AI. No API keys, no cloud costs, complete data privacy. ![Python](https://img.shields.io/badge/Python-3.14-blue) ![FastAPI](https://img.shields.io/badge/FastAPI-0.115-green) ![LangChain](https://img.shields.io/badge/LangChain-0.3-orange) ![ChromaDB](https://img.shields.io/badge/ChromaDB-1.5-purple) ![Ollama](https://img.shields.io/badge/Ollama-Local-red) ![License](https://img.shields.io/badge/License-MIT-yellow) --- ## 📋 Table of Contents - [Overview](#overview) - [Tech Stack](#tech-stack) - [Features](#features) - [Project Structure](#project-structure) - [Installation](#installation) - [Running the App](#running-the-app) - [API Endpoints](#api-endpoints) - [Supported Models](#supported-models) - [Configuration](#configuration) - [Troubleshooting](#troubleshooting) --- ## Overview Enterprise RAG System lets you upload any document (PDF, TXT, DOCX) or webpage and ask questions about it using a local LLM running entirely on your machine. No data ever leaves your computer. **Total running cost: $0/month.** --- ## Tech Stack | Layer | Technology | |-------|-----------| | LLM | Ollama (Mistral, Gemma2, Qwen2, DeepSeek) | | Embeddings | HuggingFace `all-MiniLM-L6-v2` (local) | | Vector Store | ChromaDB | | Backend API | FastAPI + Uvicorn | | Frontend UI | Streamlit | | Framework | LangChain 0.3 | | Language | Python 3.14 | --- ## Features - 📄 **Multi-format ingestion** — PDF, TXT, DOCX, URLs - 🤖 **Multi-model support** — switch between models at runtime - 🌊 **Streaming responses** — tokens appear live as generated - 💬 **Chat history** — full conversation saved in session - 📥 **Export chat** — download as TXT or JSON - 📊 **Vector store stats** — live chunk count and model info - 🗄️ **Document management** — track and clear ingested docs - 🔍 **MMR retrieval** — diverse, relevant chunk selectio