A little less work. A little more you.

Generate an SRT subtitle file locally

Turn spoken words into a timed subtitle file. Kitecut downloads a speech model and runs transcription on your device, then exports SRT subtitles alongside the result.

Choose a file to get started

Free local tools. No watermark. No account needed for local editing. Browser and device limits apply.
Updated 2026-11-03 by Kitecut.

How to use it

Choose a video with a clear spoken audio track.
Review the captions step: SRT output is enabled and burned-in captions are off. The speech model downloads when needed.
Run the plan and download the SRT from the result. Review and correct the subtitle text and timings before publishing.

A subtitle file is different from burned-in text

SRT stores text and timing separately from the picture. A compatible player or publishing tool can load it as a subtitle track. The initial plan does not put permanent text into the video's pixels.

Enable burned-in captions in the editor if the words need to be visible in the picture itself. The editor also supports VTT subtitle output.

Allow time for the model and review the words

The first transcription needs a model download. Processing time and available memory depend on your device, browser, and clip length. Background noise, music, accents, and overlapping speakers can reduce accuracy.

Automatic subtitles can mishear names and specialized language. Review both text and timing in your subtitle workflow. There is no promise of perfect recognition or support for every language.

Questions before you start

Is speech sent to an AI server to transcribe it?

Transcription runs locally after the supporting model downloads. Transcript text is shared with hosted AI only when you explicitly enable transcript sharing.

Can I use this without an AI allowance?

Local transcription and exports do not consume hosted AI requests. Optional conversations with the AI use the separate monthly allowance.