Transcribe video to text automatically

Transcript or subtitle track — decide before you export

These are the same job with two different outputs, and the one you want depends on what happens next. A transcript is a document: continuous text, no timings, useful for turning a talk into an article or making an archive searchable. A subtitle track is a file of timed cues that sits alongside the video and can be switched on and off.

Both come out of a single upload, so there is no need to choose in advance. Export the TXT for the writing and the SRT for the player.

The audio is what gets transcribed

The picture is not analysed — speech is read from the video's audio track, which is why a well-shot video with poor sound transcribes worse than a plain screen recording with a good microphone. If the footage was recorded with a separate audio device, transcribing that file directly usually gives a better result than transcribing the camera's own track.

On-screen text, slides and captions already burned into the picture are not picked up, because nothing is read from the image.

Video formats accepted

MP4, M4V, MOV, MKV, WEBM, TS and M2TS, at any size. There is no need to extract the audio first or convert anything — upload the file you already have, and your browser sends up the audio out of it rather than the whole thing. Older containers like AVI, WMV and FLV are not accepted, because that step happens in the browser and the browser cannot open them; converting one to MP4 first is enough.

What people do with the text

• Publish a subtitle track, so the video is watchable with the sound off.
• Turn a recorded talk or webinar into a written post without retyping it.
• Edit by transcript — find the sentence in the text, then cut to it in the timeline.
• Make a back catalogue searchable, which a video file otherwise is not.
• Meet accessibility requirements that ask for a text alternative.