AI Notes: A RAG-Driven AI-Powered Personal Knowledge Base

Overview

AI Notes is a self-hostable, markdown-based note-taking web application designed to provide users with a privacy-focused, AI-powered knowledge management system. The application integrates Retrieval-Augmented Generation (RAG) and semantic search capabilities, allowing users to interact with their notes using an AI chatbot that generates responses based on stored content. The system is optimized for in-browser execution using Small Language Models (SLMs) and lightweight vector databases, ensuring efficient and secure knowledge retrieval without relying on cloud-based services.

Key Features

• Markdown-Based Note-Taking: Create, edit, and organize notes using markdown syntax.
• Semantic Search: Efficiently search through notes using vector-based semantic search.
• AI-Powered Querying: Interact with your notes using an AI chatbot powered by Retrieval-Augmented Generation (RAG).
• In-Browser Execution: AI models run directly in the browser using ONNX Runtime, ensuring privacy and reducing reliance on external servers.
• Self-Hostable: Deploy the application on your own infrastructure using Docker.
• Lightweight Vector Database: Utilizes LanceDB for efficient storage and retrieval of embeddings.
• Cross-Platform Accessibility: Built as a Progressive Web App (PWA), ensuring a responsive and lightweight user experience across devices.

Technologies Used

• Frontend: Next.js
• Backend: Next.js API routes
• AI Models: Small Language Models (SLMs) optimized via distillation and quantization
• Vector Database: LanceDB
• In-Browser Inference: ONNX Runtime / Transformers.js
• Containerization: Docker
• Version Control: Git (GitHub)

Installation

Prerequisites

• Node.js (v16 or higher)
• Docker (optional, for self-hosting)
• Git

Steps

1. Clone the Repository:

   git clone https://github.com/your-username/ai-notes.git
   cd ai-notes
   

2. Install Dependencies:

   npm install
   

3. Run the Application Locally:

   npm run dev
   

   The application will be available at http://localhost:3000.

4. Self-Hosting with Docker:

   • Build the Docker image:

     docker build -t ai-notes .
     

   • Run the Docker container:

     docker run -p 3000:3000 ai-notes
     

     The application will be available at http://localhost:3000.

Usage

1. Create Notes: Use the markdown editor to create and edit notes.
2. Semantic Search: Use the search bar to find relevant notes based on semantic similarity.
3. AI Chatbot: Interact with the AI chatbot to query your notes. The chatbot will generate responses based on the content of your notes using RAG.

Project Structure

ai-notes/
├── public/              # Static assets
├── app/                 # Source code
│   ├── _components/     # Shared React components
│   ├── _hooks/          # Custom React hooks
│   ├── _providers/      # Context providers (e.g., AI, Toast, Query)
│   ├── _store/          # Zustand stores for state management
│   ├── _utils/          # Utility functions
│   ├── _workers/        # Web workers for AI tasks
│   ├── api/             # API routes for backend logic
│   ├── chat/            # Chat-related components and pages
│   ├── graph/           # Graph visualization components
│   ├── vault/           # Note management components
│   └── layout.tsx       # Application layout
├── Dockerfile           # Docker configuration
├── next.config.js       # Next.js configuration
├── package.json         # Node.js dependencies and scripts
└── README.md            # Project documentation

Contributing

Contributions are welcome! If you'd like to contribute, please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bugfix.
3. Commit your changes and push to your branch.
4. Submit a pull request with a detailed description of your changes.

Contact

For any questions or feedback, please contact:

• Oscar Rey Castro
  Email: or24aab@herts.ac.uk
  GitHub: oscarrc

---

AI Notes is a project developed as part of the BSc Computer Science program at the University of Hertfordshire.
Supervisor: Dr. Iain Werry
Academic Year: 2024/25