Minedset Capture

Minedset Capture turns whatever you are reading — or watching on YouTube — into a clean, basic file (Markdown), written directly into a folder on your computer. Click the toolbar button on an article and Minedset captures the readable text, the title, the author, the publication and the URL. On a YouTube video it captures the full transcript. Every capture can carry a short comment where you say why you grabbed it, so a note you open in six months still knows what you wanted from it. Everything runs on your own machine. The extension talks only to the Minedset application on localhost — the same computer you are sitting at. No account, no cloud service, no analytics, no telemetry. There is no server of ours for your reading to pass through, because there is no server of ours. Requires the Minedset desktop application (Windows, Python), which is free to install and free for everything on this list. The extension is the capture button; the application does the work and writes the notes. Install it first — the extension has nothing to talk to without it. Speech to text is a $5 one-time unlock inside that application — podcasts, local audio and video without captions. Nothing on the list above needs it. What it handles today: Articles, via Readability extraction — the same engine reader modes use YouTube — full transcripts, plus an in-page capture button on video pages Substack — posts and notes, including transcripts of Substack-native video when captions exist Gmail — unwraps Google's link redirects so you capture the real article, not the redirect A paste-transcript fallback for videos where captions are blocked Notes are plain Markdown with YAML frontmatter, written into a folder you choose. They open in any editor — Legend Text, or even a text editor from 1994. There is no database and no lock-in: the notes are files, and they are yours.