Skim Recap is an innovative productivity and artificial intelligence tool designed to help users efficiently digest web content. If you tend to flick through an article too quickly, Skim Recap automatically catches what you missed and hands the skipped passage back in a convenient card right beside your cursor.

Beyond catching skipped sections, the tool features a dedicated "Feynman: Gemma" function. When an unfamiliar term or complex phrase stops you in your tracks, simply select the word or phrase. The system uses the surrounding paragraph, nearby context, headings, and page title to provide a clear explanation tailored specifically to how it is used in context.

Built with privacy and performance in mind, everything runs completely locally through LiteRT-LM and WebGPU. This means you do not need to create an account, nor does it rely on any hosted LLM APIs.

Pros of Skim Recap

Recaps skipped passages instantly in a cursor-adjacent card
Feynman feature explains selected terms using local context and headings
Runs 100% locally via LiteRT-LM and WebGPU for maximum privacy
Requires no user account creation or hosted LLM API subscriptions

Cons of Skim Recap

Relies on WebGPU capabilities which may require a modern, compatible browser or hardware setup
Limited details available on mobile or non-Chrome environments based on available integration scopes

Frequently Asked Questions

What is Skim Recap?

Skim Recap is an AI-powered tool that recaps passages you skim past too quickly and provides contextual explanations for difficult terms you select.

How does Skim Recap protect my privacy?

Everything runs locally through LiteRT-LM and WebGPU. It requires no user account and does not rely on any hosted LLM APIs.

What is the Feynman feature?

The Feynman feature allows you to select any confusing word or phrase, prompting Gemma to explain its meaning using the surrounding paragraph, context, headings, and page title.