Nano-RAG - Lightweight Edge RAG Framework

Lightweight Edge RAG Framework - Intelligent Document Q&A System based on LangGraph
_{Next.js + Cloudflare Workers + D1}

Features • Quick Start • Architecture • Deployment

Live Demo →

---

Introduction

Nano-RAG is a lightweight Retrieval-Augmented Generation (RAG) application designed for individuals and small teams. It enables PDF document upload and generates high-quality Q&A responses through intelligent retrieval and LLM.

Core Features:

• 🚀 Edge Deployment - Based on Cloudflare Workers, low latency globally
• 🧠 Intelligent Retrieval - Multi-stage RAG pipeline orchestrated by LangGraph
• 💾 Serverless Storage - D1 database + vector retrieval, no extra services needed
• 🎯 Quality Assurance - Built-in hallucination detection, quality check and auto-repair

Features

Document Management

• PDF document upload and parsing
• Automatic chunking (800-1000 characters/chunk)
• Keyword extraction and document classification

Intelligent Q&A

• Multi-turn conversation support
• Streaming response output
• Source citation tracing

RAG Pipeline

Question → Relevance Check → Query Rewrite → Document Selection → Vector Retrieval → Generate Answer → Hallucination Check → Quality Check
                                      ↑________________Retry (max 2 times)________________↓

Node	Function
relevance_check	Check if question is suitable for document retrieval
rewrite_query	Optimize query statement
document_selector	Select relevant documents
retrieve	Vector similarity retrieval Top-5
generate	LLM generate answer
hallucination_check	Hallucination detection
quality_check	Quality check and repair

Tech Stack

Category	Technology
Frontend	Next.js 16, React 19, Tailwind CSS 4
Backend	Cloudflare Workers, OpenNext
Database	Cloudflare D1 (SQLite)
Orchestration	LangGraph

Quick Start

Requirements

• Node.js 22.13+

Installation

# Clone project
git clone https://github.com/your-username/nano-rag.git
cd nano-rag

# Install dependencies
npm install

# Configure environment variables
cp .dev.vars.example .dev.vars
# Edit .dev.vars, fill in OPENAI_API_KEY

# Initialize local database
npx wrangler d1 execute nano-rag-db --local --file=./migrations/0001_init.sql

Development

# Local preview (Cloudflare Workers runtime, with D1)
npm run preview
# Visit http://localhost:8787

# Pure dev mode (frontend only, no database)
npm run dev
# Visit http://localhost:3000

Deployment

# Create production database
npx wrangler d1 create nano-rag-production-db

# Update database_id in wrangler.toml

# Execute database migration
npx wrangler d1 execute nano-rag-production-db --remote --file=./migrations/0001_init.sql

# Configure Secrets
npx wrangler secret put OPENAI_API_KEY

# Deploy
npm run deploy

For detailed deployment guide, see DEPLOYMENT.md.

_{Built with ❤️ using Next.js, Cloudflare Workers, and LangGraph}