summarize
Point at any URL/YouTube/Podcast or file. Get the gist. CLI and Chrome Extension.
Summarize is a tool that takes a URL, a file, or a media source and produces a plain-English summary using an AI language model. It comes in two forms: a command-line tool you run in a terminal, and a Chrome (or Firefox) side panel extension that adds a summary button to your browser. Both share the same underlying code and can use a variety of AI providers including OpenAI, Claude, Gemini, and free models available through OpenRouter.
The CLI accepts web pages, PDFs, images, audio files, video files, YouTube links, podcast RSS feeds, and piped text from other programs. For YouTube and local video files it can also extract slides by taking screenshots at regular intervals, running text recognition on those frames, and presenting them as timestamped cards alongside the summary. Audio content is transcribed first using whichever transcription service is configured, then summarized.
The browser extension adds a side panel that summarizes the current tab on demand or automatically as you navigate between pages. Because heavy media processing (downloading videos, running transcription) cannot happen inside a browser extension, the extension communicates with a small background service you install on your machine. That service runs locally, restarts automatically on login, and only accepts connections from the browser that paired with it using a shared token.
Installation is via npm or Homebrew. Slide extraction requires ffmpeg and yt-dlp to be installed separately; OCR requires Tesseract. The tool shows cost estimates and timing information alongside summaries so users can see what the AI call cost. Output length can be configured, and a caching layer avoids re-fetching content that was recently summarized. The project is written in TypeScript and released under an open-source license.