Subtitles

Generate subtitles from any audio or video

Transcribe once, then export the same result as SRT, VTT or ASS, with sentence breaks that follow the meaning.

Direct answer

EdgeSpeak creates subtitles locally: transcription, semantic segmentation and word-level alignment, then export to SRT, VTT or ASS.

Sentence breaks that read well

Turn on semantic segmentation and lines break at natural boundaries instead of at a fixed length.

Timing you can trust

Alignment refines each word's start and end, so subtitles stay in step with the voice.

Pick the format your editor wants

SRT and VTT work in almost every player and editor. ASS adds styling and karaoke highlighting.

Batch it from a script

Run the same pipeline from the CLI when you have many files.

edgespeak-cli transcribe talk.mp4 -o talk.json

FAQ

Which subtitle formats can I export?

SRT, VTT and ASS, plus JSON and plain text for your own tools.

Can I fix errors before exporting?

Yes. Edit any line in the transcript while it plays, then export.

Can I translate the subtitles?

Yes. EdgeSpeak translates transcripts on your computer with an on-device model.

Do I need to upload the video?

No. The video is processed on your computer.

Evidence links