Skip to content

Wikipedia fact checked by AI and RAG, then brought to you by AI.

License

Notifications You must be signed in to change notification settings

MyMel2001/Nodeapedia

Repository files navigation

Nodeapedia 🌐🧠

Nodeapedia is an AI-powered Wikipedia reimagining. It takes traditional Wikipedia articles and enhances them using a Retrieval-Augmented Generation (RAG) engine. Instead of just reading a static entry, you get an AI-generated synthesis that is cross-referenced against live web search results to identify discrepancies, missing context, and interesting additional facts.

🚀 How It Works

  1. Wikipedia Fetch: Retrieves the core topic information from Wikipedia's REST API.
  2. Web Verification: Uses a LibreY instance to perform a privacy-respecting search for the latest information on the topic.
  3. Content Scraping: Scrapes the top web results to gather deep context (not just snippets).
  4. AI Fact-Check: A smaller LLM (e.g., qwen3:4b) compares the Wikipedia data with the web results to find inconsistencies.
  5. Synthesis: A larger LLM (e.g., deepseek-r1:14b) writes a concise, easy-to-read summary incorporating the fact-check findings.
  6. Sources: All original sources (Wikipedia + Web results) are provided to the user for transparency.

🛠 Self-Hosting Guide

Hosting your own instance of Nodeapedia requires a local LLM runner (Ollama) and a Node.js environment.

Prerequisites

  • Node.js (v18+ recommended)
  • Ollama (Running locally or on a reachable server)
  • Access to a LibreY instance (defaults to search.sparksammy.com)

1. Setup Models

Ensure you have the required models pulled in Ollama:

ollama pull gemma3:4b
ollama pull sparksammy/ministral-3-14b-unsloth:small-hotfixed

2. Installation

Clone the repository and install dependencies:

git clone https://github.com/MyMel2001/Nodeapedia.git
cd Nodeapedia
npm install

3. Configuration

Create a .env file in the root directory (or copy .env.example):

# Ollama Configuration
OLLAMA_BASE_URL=http://localhost:11434/v1
MODEL_FACT_CHECK=qwen3:4b
MODEL_SUMMARY=deepseek-r1:14b

# Search Configuration (LibreY)
LIBREY_BASE_URL=https://search.sparksammy.com

# Server Configuration
PORT=6107
NODE_ENV=development

4. Build and Run

Generate the CSS (if modified) and start the server:

# Build Tailwind CSS
npm run build:css

# Start the application
npm start

Your instance will be running at http://localhost:6107.

📜 License

This project is licensed under the SPL-R5 License.

About

Wikipedia fact checked by AI and RAG, then brought to you by AI.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors