mwalker22/TMD-Open-Source-RAG

TMD Open Source RAG

A Retrieval-Augmented Generation (RAG) app using open-source models, FastAPI, and Vite/React.# 🌽 Field of Dream Deploys

_“If you deploy it, they can use it.”_

Welcome to field-of-dream-deploys, a full-stack Retrieval-Augmented Generation (RAG) application powered by open-source models deployed via Hugging Face Inference Endpoints.

This project showcases how to:

• —Load and embed real-world documents (like essays and engineering philosophies)
• —Persist them in a vector database
• —Ask questions through a conversational React frontend
• —Route responses through your own LLM and embedding endpoints — no OpenAI required

Inspired by Field of Dreams, this app turns deployed endpoints into usable interfaces for real-time insight and discovery.

Built with:

• —LangChain
• —FastAPI
• —React
• —Qdrant
• —Hugging Face Inference Endpoints

Step onto the field. Deploy the dream. Let the questions come.

Backend Template (/backend)

This backend provides a reusable scaffold for building LLM-powered applications with FastAPI.

Features

• —Modular routing with /api
• —Prompt manager and role-based prompt composition
• —Plug-and-play vector DB with FAISS
• —Easily swappable embedding model