Convert audio to subtitles

Subtitles from audio, without a video file

Speech recognition works on the audio track, so a video is never actually required — the picture contributes nothing to the transcript. Uploading audio directly skips a step and produces a smaller file to transfer.

This matters most for podcasts, recorded interviews, lectures and voice notes, where a video version either does not exist or is added later. The subtitle file generated from audio will drop straight onto a video of the same recording, because the timings refer to the audio either way.

Supported audio formats

Common formats all work: mp3, wav, m4a, aac, flac, ogg and opus, along with the audio track of any video file you upload. Files are converted to a consistent format internally before recognition, so the source encoding does not affect accuracy.

Recording quality does affect accuracy, considerably. A close microphone in a quiet room transcribes far better than a phone across a table, and separating speakers onto their own tracks — where you have that option — helps more than any other single thing.

Podcasts and long recordings

There is no length limit, and long files are not slower in the way you would expect: recordings are segmented and transcribed in parallel, so a two-hour episode finishes in broadly the same time as a ten-minute one.

Cost is one credit per minute rounded up per file, so a 45-minute episode costs 45 credits out of the 100 that $10 buys.

Transcript or subtitle file

Export as TXT for a plain transcript with no timecodes — show notes, articles, searchable archives — or as SRT and VTT for a timed track. The same job produces all three, and can also translate the result into any of the 76 supported languages.