Skiff - save to Claude, developed by Skiff, is a Chrome extension that captures the web pages a signed-in user sees and forwards them to Anthropic’s Claude for analysis. It converts cluttered pages into clean text, extracts YouTube transcripts, and supports one-tap clipping for rapid capture. Built for AI researchers, journalists, and power users, the tool helps incorporate live social threads and long-form articles into Claude-driven research workflows without relying on external scrapers.

What is Skiff used for?

Skiff converts visible, logged-in web content into machine-readable text for Claude. The extension operates inside the browser session you’re already using, so it captures X threads, Reddit discussions, paywalled articles, and video transcripts as you view them. A 60-second setup and a single toolbar action let users save pages into a text library that Claude can access during chat sessions.

How does Skiff handle pages that block AI crawlers?

It bypasses server-side scraper blocks by capturing the page from your authenticated browser view. Rather than attempting background crawling, the extension reads the DOM you see and strips ads, navigation, and scripts to produce structured text. Supported capture types mentioned include social threads, long-form articles, and YouTube transcripts, which the tool formats to reduce noise before delivery to Claude.

Does Skiff reduce privacy risk while sharing content with Claude?

Privacy is an explicit design focus, but the workflow requires trade-offs tied to logged-in access. The team’s background in end-to-end encryption informs the extension’s emphasis on clean data handling. At the same time, capturing paywalled or account-only content depends on your active session, and saved text is sent to a Claude-readable library, so users must accept that captured material is made available to the AI environment.

How does Skiff fit into a Claude-based research workflow?

Skiff integrates directly with Claude to make live web material immediately usable for analysis. The direct-to-library transfer means researchers and journalists can gather source text from diverse pages without crafting manual copies. The extension is aimed at power users who need rapid, repeatable capture; setup is brief and the result is structured input that Claude can reference during conversations or longer analysis tasks.

Skiff is a practical tool for Claude-centric research, with a clear trade-off.

Skiff is well suited to researchers and power users who need to bring what they see in-browser into Claude for immediate analysis. The main trade-off is reliance on desktop browser sessions to reach account-only and paywalled material, which places responsibility for logged-in access and data sharing squarely with the user. For focused Claude workflows, the extension shortens the gap between web research and model input.