Tracy - Reading Companion

Our users say they actually read articles and papers after asking Tracy for a quick summary! Tracy is a hands-free voice assistant that lives on every webpage you visit and PDF you read. Simply ask a question naturally, and Tracy explains it — no typing, copy-pasting or switching tabs. ---------------HOW IT WORKS--------------- • Websites: Point your cursor at what you want explained & simply ask • PDFs: Tracy sees what you're seeing. Simply ask. It's that simple! -------------------FEATURES------------------- • Works on all webpages/PDFs— articles, documents, papers, Wikipedia, blogs • Sees what's on your screen via a customizable viewport overlay • Understands the full page context automatically • Speaks answers aloud — fully conversational, no reading required • Supports multiple AI models: ChatGPT, Claude, Gemini • Customizable viewport size, shape, transparency, and contrast • Multi-language support • Send voice feedback to the developer ------TWO WAYS TO USE TRACY------ Tracy Pro — $5/month • No API keys, no setup — everything just works • All AI models handled for you • Text-to-speech included • Priority support • Subscribe from within the extension Bring Your Own Key — Free • Use your own API key from OpenAI, Anthropic, or Google • Enter your key in the extension popup and you're ready • Your key stays in your browser— we never store, transmit, or see it --------------------PRIVACY-------------------- • Only allowed on a website/PDF after you explicitly give permission to run • BYOK: API calls go directly from your browser to the provider • Pro: Queries proxied through our secure server— we don't log/store your queries • Tracy collects installation-linked usage and error counts, provider/model details, request status, timing, and token or speech-character counts to improve reliability and estimate provider costs. Country code is also recorded. Records may be linked to an email you provide or your Pro license. These operational records exclude API keys, questions, answers, page content, screenshots, and reading URLs.