How long does it take? Captions vs. Whisper?

Time

About 45 seconds on average, and usually under a minute. Factors: video length, whether captions already exist, and current load.

Captions first, transcription fallback

• Has captions: used directly — fastest and cheapest.
• No captions: speech-to-text (Whisper) converts the audio to text, taking a bit more time and cost.

Captions that YouTube produced by auto-translating or auto-dubbing the video don't count as existing captions. For example, when a Chinese video only has auto-generated English captions, we transcribe the original audio instead, so names and terms stay accurate. These videos are priced as videos without captions.

Some videos have incomplete captions: only the first few lines, or mostly blank. When the captions cover less than half of the video, we transcribe the audio instead, so the whole video is covered. This takes a little longer, and you are not charged extra for the switch to transcription.

The result page shows whether this video used "captions" or "transcription".