Export workflow · Updated November 3, 2026
Save a video transcript as Markdown
AnyVideo's Chrome extension exports the spoken content of a supported public video to a .md file. The file starts with a title and source details, followed by the transcript, so you can keep the recording attached to your notes.
What the Markdown file contains
- A level-one heading with the video title when available.
- The source URL and transcription date.
- The method: platform subtitles or speech recognition.
- Approximate video duration when available, then the transcript body.
The example below uses a short passage from a recorded 31:39 browser-audio test. The Markdown header illustrates the current export format; the body is an excerpt, not the full recording.
# The Future of AI Agents with Andrew Ng | Interrupt 26 - Source: https://www.youtube.com/watch?v=OaRhpwz_TGM - Transcribed: 2026-10-28 - Method: speech recognition - Duration: ~32 min Hey, everyone. Thank you for being here, Andrew. It's great to have you back
Download Markdown excerpt · Download text excerpt
From video link to your Obsidian vault
- Install AnyVideo in Chrome. Submit a supported public video link in the side panel. Follow the full transcription walkthrough if this is your first job.
- When the transcript is ready, choose Download MD. Save it before the server copy expires after 7 days.
- Move the .md file into your Obsidian vault folder, or open it with another Markdown editor. Obsidian is a separate app; AnyVideo does not connect to your vault or sync notes automatically.
- Review names, numbers and quotations against the source video. Add your own headings, highlights and links where useful.
- Keep your summary separate from the transcript so you can distinguish the speaker's words from your interpretation.
Markdown or plain text?
Choose Markdown when you want the title, source and method alongside the words. Choose TXT when you only need the transcript body for search, copying or editing. Both are editable text files.
A transcript is not a summary
The export does not generate topic headings, takeaways, a translation, speaker labels or timestamped subtitles. It also cannot recover unspoken text from slides. If you use a separate AI tool to summarize the file, check that tool's data policy before uploading it and verify its claims against the original.
Before you cite a passage
Platform captions can contain errors; speech recognition can mishear unfamiliar names and overlapping voices. Keep the source link, replay important passages and correct the text yourself. A transcript alone does not establish that a quotation is accurate.
Tested sources and restrictions · Plans and top-ups · Audio handling and retention