Web Clipper for pages, AI chats, and Markdown notes
Right-click any page, AI chat, or selected text and turn it into clean Markdown in Mark It Down. Preview, trim, save to Inbox, and edit in Side Panel.
Right-click, preview, then keep the part worth rewriting.
Web Clipper is the Entry surface for web pages and AI output. It turns selected text or full pages into ordinary Markdown notes, then hands them to the same local editor as everything else.
- Page and selection clipping — Save the current page or selected text as clean Markdown, with the source URL preserved.
- AI chat extraction — Capture useful output from Claude, ChatGPT, Grok, and Gemini without adding AI writing inside the editor.
- Side Panel editing — Open saved clips next to the page or chat they came from, so cleanup starts in context.
Three steps from right-click to an editable note.
Clipping usually takes the same three steps: pick an action from the context menu, check the result in a preview, and keep it as an ordinary note in Inbox.
- 1. Pick from the right-click menu — Four save actions: selected text, the full page, a YouTube transcript, or an AI response digest. Three copy actions skip saving entirely — selection as Markdown, page link as Markdown, and a selection formatted for pasting into an LLM.
- 2. Shape it in the preview — The Side Panel opens with a preview of the extracted Markdown. Include or exclude sections one by one, or switch the extraction mode between article, full page, and CSS selector. If extraction fails, the Side Panel can keep what it has so you can retry, and a save action can fall back to plain text instead — but neither fallback covers every failure path, so a clip can still be lost.
- 3. Save to Inbox — Clips always land in Inbox — no filing decision at save time. Site details like the title and URL are stored with the clip and can be pulled into frontmatter later. The note stays open next to its source page, so rewriting starts immediately.
Markdown that keeps structure instead of flattening it.
The clipper chooses extraction paths for the page shape, then keeps the result readable as plain Markdown. The goal is not bulk ingestion; it is a clean handoff into your own words.
- Readable article structure — Headings, lists, code, tables, and links stay editable after capture.
- Specialized extractors — Social threads, publishing sites, YouTube transcripts, Mermaid blocks, and Japanese ruby annotations have dedicated handling.
- Local review step — Preview and clip filters let you trim the result before it becomes a note.
Dedicated extractors where generic extraction falls apart.
Most pages go through general article extraction. For page shapes that flatten badly, the clipper switches to dedicated extractors — reaching into Shadow DOM and same-origin iframes when the content lives there.
- AI chats — Dedicated extraction for Claude, ChatGPT, Gemini, Grok, and Perplexity conversations — including UIs rendered inside Shadow DOM, like Gemini.
- Comment threads — Hacker News, Reddit, and GitHub issues, pull requests, and discussions come through as readable threads.
- Social posts — X threads, LinkedIn, Bluesky, and Threads.
- Publishing sites — Medium, Discourse forums, Zenn, Qiita, and Hatena.
- And beyond — YouTube transcripts, automatic llms.txt detection when a site publishes one, Mermaid diagrams kept as diagrams, and Japanese ruby annotations converted to Aozora-style 《reading》 notation.
Clip less. Keep what matters.
Questions about permissions or a page that will not clip? See the FAQ or Troubleshooting.