MarkSense

MarkSense turns your browser into an intelligent reading and study assistant. It summarizes, translates, and intelligently highlights web content, with all AI processing running locally on-device via Chrome's built-in Gemini Nano APIs—no API keys, no network calls, and zero data sharing.

Core features include one-click webpage summaries with adjustable tone and length, automatic translation of summaries into a chosen language, keyword extraction, AI-generated related-article cards, and interactive highlighting to flag important or confusing passages.

Built with Manifest V3, the Side Panel API, and Chrome's AI SDK (Summarizer, Prompt, and Translator APIs), using Mozilla Readability for clean text extraction and chrome.storage.local for persistent highlights. Key challenges included working around experimental API availability, chunking long articles past the summarizer's input limit, and re-anchoring highlights after the DOM changes on page reload.