Captions with word-level timing.

The engine hears the audio and times every word. Thirty studio styles render live in the editor, and the export burns in exactly what you previewed.

Timed to the word, not the line

Each word carries its own start and end. That's what makes the highlight styles feel alive, and what makes editing safe: change a word, and its timing stays put.

A transcript you can actually edit

Captions live on their own timeline track and in a transcript panel. Double-click a word to fix a name, split a line at the playhead, merge two lines, drag edges to retime; every change is undoable.

Thirty styles with a point of view

From quiet cinema lines to full karaoke highlight. The samples below are the real styles rendering live in your browser, the same code path the export uses.

Quiet lines for cinema
glass
barely there at all
whisper
The word lands on time
glide2
Krijo një herë, fol kudo
backdrop
一度作れば どこでも話せる
vinta
Cada palabra a tiempo
ali
soft but certain
kitty
LOUD AND PRECISE
vegas
Momentum in every beat
hustle

Questions

How accurate is the transcription?
Strong on clear speech in 164 recognized languages, with word-level timestamps. Names and jargon are the usual weak spot of any engine, which is why the transcript editor makes them one double-click to fix.
Can I move where captions appear?
Yes: top, center, or bottom placement, with styles that keep their look at any position.
Are captions burned into the export?
Yes. The export renders your captions in the chosen style directly into the video, identical to the editor preview.