November 9, 2026 · 4 min read
How to Transcribe a YouTube Video That Has No Captions
In short
If a YouTube video has no captions, its transcript has to come from the audio. Check for hidden caption tracks first, then use AI speech-to-text through an API or run Whisper yourself on audio you have the right to use.
When YouTube shows no Show transcript button, the video usually has no caption track at all: captions were turned off, the video is too new for automatic captions, or the language isn't supported. Then the only way to get text is to transcribe the audio with a speech-to-text model.
1. Make sure there really are no captions
Open the video on a computer, expand the description, and look for Show transcript, or click the CC button and check Settings → Subtitles/CC. Some videos have captions only in another language, and some auto-captions appear a few hours after upload.
The free YouTube video stats tool shows whether a video has captions and in which languages, and the transcript generator lets you switch between every track a video has.
2. Use an API that falls back to AI transcription
TranscriptYT's YouTube Transcript API returns the caption track when one exists and otherwise transcribes the audio with Whisper large-v3-turbo, in the same request and response shape. The response's source field says which one you got, so you always know whether the text came from captions or from speech-to-text.
It is available on paid plans and credit packs and costs 1 credit per started minute of audio (a 10-minute video is 10 credits); failed requests are never billed, and a transcript someone already generated is served from cache for 1 credit. Pass language to hint the spoken language, which improves accuracy for non-English videos.
curl "https://transcript-yt.com/v1/transcript?url=https://youtu.be/VIDEO_ID&format=text" \ -H "Authorization: Bearer ts_live_..."
3. Or run Whisper yourself
If you have the right to use the audio, for example your own video or one you have permission to process, you can download it and run OpenAI's open-source Whisper locally. It is free but needs a capable computer or GPU, and you handle the audio download, long-file splitting, and timestamps yourself.
For more than a few videos, or from a server where YouTube blocks downloads, an API is usually cheaper than the engineering time. See YouTube captions vs. Whisper for the trade-offs.
Videos that can't be transcribed
Private, members-only, and removed videos can't be transcribed by any outside tool. Some region-locked videos are retried from another country automatically, and made-for-kids videos currently can't be downloaded for AI transcription.
Common questions
Why does a YouTube video have no transcript?
The uploader turned captions off, YouTube hasn't generated automatic captions yet (new or long videos can take a while), the spoken language isn't supported for auto-captions, or the video has little or no speech.
Can I get a transcript of a YouTube video with no captions for free?
Free caption tools, including TranscriptYT's free tool, only read existing caption tracks. For a video with none, you can run the open-source Whisper model on your own computer, or use an API: TranscriptYT's paid plans transcribe the audio at 1 credit per minute, from $3 a month for 1,000 credits.
How accurate is AI transcription of YouTube audio?
Whisper large-v3-turbo is accurate on clear speech in major languages and weaker on heavy background music, overlapping speakers, and rare names.
How long can the video be?
Through the TranscriptYT API, videos up to about two hours are transcribed in a single request; a two-hour video takes roughly two minutes.