Katto
Transcription

The transcript of your video, in the format you need

Katto transcribes every video it clips, and the text is yours to take away: an SRT file for each clip, the edited text of a clip as plain text, or the full transcript of the source video with timestamps through the API. This page is about getting the words out. If you want captions burned into the video, that is a different page.

99
languages transcribed

The language of your recording is detected, or you can set it before processing.

41
timed word by word

The other languages are timed sentence by sentence. Both give correct text; only the timing precision differs.

Katto results page: a ranked clip with its score and its timestamped transcript
On the results page, each clip shows its transcript with timestamps and the pauses between sentences.

Three ways to get the text

FormatWhereWhat it contains
SRTResults page, one file per clipSubtitle lines with start and end times, ready for YouTube, a video editor or a player.
Plain textEditor, per clipThe clip's words after your corrections, to copy or download as a .txt file.
JSONAPI (GET /v1/jobs/{id}/transcript) and MCPThe whole source video: language and timestamped segments, for your own tools or an AI agent.

Names spelled right

Transcription fails first on proper nouns: a guest's name, a product, a place. A correction pass runs over them before you see the text, and anything left is editable in the clip's transcript. Exports use the corrected text.

What it is not

Katto is not a standalone transcription service. The transcript comes with a clipping job, so a video counts once against your monthly quota, and you get the clips as well as the text. If all you need is a transcript of hundreds of hours of audio, a dedicated transcription tool will suit you better.

Questions

Can I download the transcript of the whole video?

Through the API and the MCP server, yes: the full source transcript with timestamps, as JSON. In the app, the text is exported clip by clip, as SRT or plain text.

What does 41 word-by-word languages mean?

In 41 languages, every word gets its own timestamp. In the other 58, timing is per sentence. The text is accurate in all 99.

Does a transcript use my quota?

The transcript is part of clipping a video, which counts as one video of your monthly quota. There is no separate charge for the text.

Can I translate the transcript?

Translating a clip's subtitles into another language is part of the Creator plan.

Get the words out of your next video

Paste a link or upload a file, and download the transcript alongside your clips.