EasyRag

A lightweight Retrieval-Augmented Generation (RAG) system with a web interface. Upload documents, ask questions, and retrieve semantically relevant passages.

Features

• 📄 Upload TXT, PDF, DOCX files via drag-and-drop
• 🧩 Automatic chunking with sentence-aware break points
• 🔍 Semantic search powered by SentenceTransformers (all-MiniLM-L6-v2)
• 🗄️ Vector storage with sqlite-vec — no external database needed
• 🖥️ Dark-themed web UI

Quick Start

git clone https://github.com/Ratman463/EasyRag.git
cd EasyRag
pip install -r requirements.txt
python main.py

Open http://localhost:8000 in your browser.

SentenceTransformers automatically downloads the embedding model on first run — no manual setup needed.

Configuration

All settings are in config.py:

Parameter	Default	Description
EMBEDDING_MODEL	sentence-transformers/all-MiniLM-L6-v2	SentenceTransformers model name
EMBEDDING_DIMENSION	384	Vector dimension (must match model)
CHUNK_SIZE	500	Max characters per chunk
CHUNK_OVERLAP	50	Overlap between chunks
TOP_K_RESULTS	5	Default number of query results
MAX_UPLOAD_SIZE	50 MB	Max upload file size
DATABASE_PATH	data/rag.db	SQLite database path

Project Structure

EasyRag/
├── main.py          # FastAPI app & API endpoints
├── config.py        # Configuration
├── database.py      # sqlite-vec vector database
├── embedding.py     # SentenceTransformers embedding & chunking
├── requirements.txt
├── data/rag.db      # Database (auto-created)
├── uploads/         # Uploaded files
└── static/          # Web UI (HTML/CSS/JS)