Project Overview
Readie is a universal reading copilot built as a Chrome extension that works on any text displayed on your screen—research papers, PDFs, web articles, or technical documentation. By drawing a selection box over any section of content, you can ask natural language questions and get instant explanations powered by OpenAI's GPT-4 vision model, all without leaving the page.
The extension is designed for privacy and simplicity: API keys are stored locally, requests are made directly from the client to OpenAI, and no backend infrastructure is required. Readie boosts comprehension and productivity for students, researchers, and engineers tackling dense material across any website.
Key Features
- Screen Selection Overlay: Press
Alt/Option + Shift + Rto drag a selection anywhere on the page and start asking questions. - GPT-4 Vision: Leverages OpenAI's latest vision-capable model to understand screenshots and complex layouts.
- Privacy First: Keeps API keys on-device and avoids proxy servers or shared infrastructure.
- Zero Backend: Entirely client-side for easy setup and rapid response times.
- Works Everywhere: Compatible with any webpage content, including embedded PDFs and documentation portals.
Architecture & Stack
The extension is composed of a Chrome content script that captures user selections, a popup UI for configuration, and a background service worker that handles authenticated calls to the OpenAI GPT-4 Vision API. The implementation emphasizes a light footprint—HTML, CSS, and vanilla JavaScript—so it can be installed directly from source without build tooling.