Speed Reader is an open-source, AI-powered web app that helps you read, learn, and retain information from long texts, lectures, and documents. It combines a rapid reading interface with persistent session management and local LLM (Large Language Model) integration for personalized recaps, checkpoints, and interactive learning.

• Speed Reading UI: Read any text at your chosen pace, with word-by-word or chunked display.
• Session Persistence: Automatically saves your reading position, answered questions, and progress for every text—resume anytime, even after closing the browser.
• AI-Powered Checkpoints: At intervals, the app asks comprehension questions and provides instant feedback using a local LLM (e.g., llama.cpp, Qwen3.5-27B).
• Seamless Resumption: When you return, the app bridges your last session with a smart AI-generated recap, so you never lose context.
• Local-First & Private: All session data is stored locally in your browser (IndexedDB). No cloud required. LLM runs locally or on your own server.
• Custom LLM Provider: Easily configure your own OpenAI-compatible LLM endpoint (supports llama.cpp, OpenRouter, etc.).
• Robust Error Handling: Persistent loading overlays and clear feedback ensure the UI never gets stuck, even with slow or failing AI responses.
• Modern, Responsive UI: Built with React, designed for clarity and focus.
• Tested & Reliable: 80+ unit tests, clean build, and robust async flows.

git clone https://github.com/yourusername/speed-reader.git
cd speed-reader

npm install

npm start

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

Speed Reader is designed to work with any OpenAI-compatible LLM API (e.g., llama.cpp, OpenRouter, local server). For best results:

1. Run a local LLM server (e.g., llama.cpp) with the /v1/chat/completions endpoint enabled.
2. Configure the provider URL in the app's Home screen (auto-filled for common setups).
3. Choose your model (Qwen3.5-27B, Llama-2, etc.) and adjust max tokens as needed.

Note: All AI processing happens locally or on your server—your data never leaves your machine.

public/
  index.html
  demo.gif
src/
  App.js            # Main app entry
  index.js          # React entry point
  db/sessions.js    # IndexedDB session persistence
  hooks/useSpeedReader.js # Reading/session logic
  components/
    ReadingScreen.js
    HomeScreen.js
    ResumeOverlay.js
  api/llm.js        # LLM API integration
  __tests__/
    api/llm.test.js # Unit tests
package.json
README.md

• Reading Loop: The app displays text at your chosen speed, pausing at checkpoints for AI-generated questions.
• Session Save/Resume: Progress, answers, and text are saved in IndexedDB. Resume anytime, even after a browser restart.
• AI Recaps: When resuming, the LLM summarizes your last session and bridges to the new one, so you never lose context.
• Error Handling: Persistent overlays and robust async flows ensure the UI never gets stuck, even if the LLM is slow or fails.

• Add New LLM Providers: Point the provider URL to any OpenAI-compatible endpoint.
• Customize Checkpoints: Tweak the checkpoint logic in useSpeedReader.js for different question types or intervals.
• UI Themes: Style the app via CSS or add new components for different reading modes.
• Analytics & Export: All data is local, but you can add export/import features for your sessions.

Contributions are welcome! Please open issues or pull requests for bugs, features, or improvements.

1. Fork the repo
2. Create a feature branch (git checkout -b feature/your-feature)
3. Commit your changes
4. Push and open a PR

MIT License. See LICENSE for details.

• llama.cpp for blazing-fast local LLMs
• idb for IndexedDB convenience
• OpenRouter for API compatibility inspiration
• All open-source contributors and the AI community

Reading is the foundation of learning, but long texts and lectures can be overwhelming. Speed Reader helps you:

• Stay focused and engaged
• Retain more with active checkpoints
• Learn at your own pace
• Own your data and privacy
• Leverage the power of local AI

Open source, private, and built for curious minds. Happy reading!