Transcribe YouTube videos with Whisper — even without captions
Most "YouTube transcript" tools just scrape captions that someone already added — so they fail the moment a video has none. Txtify pulls the audio and runs real Whisper speech recognition, so you get an accurate transcript from any video, captions or not. Free, no account, and self-hostable.
Caption scrapers vs real transcription
| Txtify (Whisper) | Caption scrapers | |
|---|---|---|
| Works when the video has no captions | Yes | No |
| Accuracy source | Transcribes the audio | Only as good as existing captions |
| Subtitle export (SRT/VTT/SBV) | Yes | Varies |
| Translation | Yes (DeepL, optional) | Usually no |
| Self-hostable / private | Yes | No |
| Account required | No | Often |
How it works
- Paste the URL. Txtify fetches the audio track from the video.
- Pick a Whisper model. Trade speed for accuracy depending on the video.
- Get your transcript. Export TXT for a clean read, or SRT/VTT/SBV for subtitles — optionally translated into 30+ languages.
About limits: the public demo at txtify.lkmeta.com caps YouTube videos at 15 minutes by default so the shared server stays responsive. When you self-host, you raise that limit yourself — long lectures and podcasts are only bounded by your own hardware.
Good for
- Lectures, talks and podcasts with no official captions
- Turning a video into a searchable, quotable text document
- Subtitling your own uploads before publishing
- Translating a foreign-language video into subtitles you can read
FAQ
Does it need the video to have captions?
No — Txtify transcribes the audio itself, so captions aren't required.
Can I get subtitles, not just text?
Yes — export SRT, VTT or SBV. See generate SRT subtitles.
Is my activity private?
Self-host and the whole process runs on your machine. The only optional external call is DeepL for translation.
Related: Otter.ai open-source alternative · Best self-hosted transcription tools · Whishper vs Scriberr vs Txtify · Generate SRT subtitles · Private, offline transcription