RAGForge
Chat with Your Documents.
Fully Local. Fully Private.

A professional Retrieval-Augmented Generation (RAG) system built with Streamlit, FAISS, and Ollama — no cloud, no data leaks.

100% Local
FAISS Powered
LLM-Agnostic
Modular Architecture
RAGForge Terminal

$ streamlit run app.py

🚀 Launching RAGForge...

📚 Loading knowledge base: research_papers

🧠 Model: llama3.2 via Ollama

✅ FAISS index loaded: 1,247 vectors

You: What are the key findings about neural networks?

RAGForge: Based on your documents, the key findings include...

Why RAGForge?

Better than ChatGPT + uploads. Here's why developers and researchers choose us.

Privacy-First

Your documents never leave your machine. No API calls, no telemetry, no vendor lock-in.

Context-Aware Chat

Ask complex questions and get grounded answers powered by vector similarity, not keyword matching.

Multi-Knowledge Bases

Organize documents under isolated titles and chat with them independently or together.

Developer-Friendly

Clean architecture, readable code, and a full learning phase for understanding RAG from scratch.

How It Works

From document to intelligent conversation in four simple steps.

Step 01

Upload Documents

Drop PDFs, text files, or paste content directly

Step 02

Chunk & Embed

Documents split into chunks, embedded via Sentence-Transformers

Step 03

Store in FAISS

Vectors indexed for lightning-fast similarity search

Step 04

Chat & Retrieve

Query with Ollama, get grounded answers from your docs

🔍 Want the full picture? See the architecture below →

Who is RAGForge for?

Whether you're a student, developer, or professional — RAGForge adapts to your needs.

Students & Researchers

  • Chat with PDFs, notes, and books
  • Build private study copilots
  • Extract insights without cloud uploads

Developers

  • Query codebases intelligently
  • Offline documentation assistant
  • Learn RAG internals hands-on

Privacy-Conscious Users

  • No cloud dependency whatsoever
  • Safe for sensitive documents
  • Full control over your data

Knowledge Workers

  • Personal knowledge vault
  • Long-term memory assistant
  • Organize multiple knowledge bases

Get Started in Minutes

No complex setup. Just clone, install, and run.

0Prerequisites

Python 3.10+

Install →

Ollama installed

Install →

Model pulled (e.g., llama3.2)

ollama pull llama3.2 # or any other model

Quick Install

Clone the repository
$git clone https://github.com/koffandaff/RAGForge.git
Install dependencies
$cd RAGForge && pip install -r requirements.txt
Launch the application
$streamlit run app.py
Access in browser
$# Open http://localhost:8501

That's it! You're ready to chat with your documents. 🎉

Built to Teach

This project is designed not just to work — but to help you understand RAG from the ground up.

LearningPhase/

Educational modules included

Every core RAG concept is broken down into isolated, heavily-commented Python modules. Perfect for students, educators, and anyone wanting to understand the internals.

RAGForge/
├── LearningPhase/
│   ├── 01_chunking.py      # Document splitting
│   ├── 02_embeddings.py    # Vector generation
│   ├── 03_faiss_basics.py  # Index operations
│   └── 04_rag_pipeline.py  # Full RAG flow
├── app.py                  # Main Streamlit app
└── core/                   # Production modules

Chunking & Preprocessing

Learn how documents are split into semantic chunks for optimal retrieval.

Embedding Generation

Understand how Sentence-Transformers convert text to vectors.

Vector Search Fundamentals

Master FAISS indexing and similarity search algorithms.

📖 Each module is self-contained and runnable independently.

Built for Trust

Transparency and reliability at every level.

Battle-tested libraries

Built on FAISS, Sentence-Transformers, and Ollama

Works fully offline

No internet required after initial setup

MIT Licensed

Free to use, modify, and distribute

No external APIs

Zero data leaves your machine

Open source

Fully transparent, auditable code

Community driven

Open to contributions and feedback

Ready to Get Started?

Explore the code, star the repo, and start chatting with your documents today.