Give each word
its moment.
Keep the phrase still. Let the spoken word light up. Write short lines, prepare a voiced cut, then fine-tune any word beside the preview.
5.96 seconds · 720 × 1280 · 24 fps · three complete local voices · twelve timed words.
Preview, then make it yours.
In Film, open Make a narrated cut and choose Follow voice. Review the prepared film before using it. Every phrase and word time stays editable in Typography. Changing the wording clears its old speech timing.
Timing comes from the generated voice model. It is estimated, so review it against the audio and adjust as needed. No additional runtime model is downloaded for this style. Original footage and supplied WAVs are in the complete source archives.
Review notes and timing evidence
This is a real native UI export. Each title uses the timing returned with its exact Kokoro WAV. All 143 frames, complete voices, source hashes and embedded project were checked; selected frames cover every highlighted word.
An optional open-weight CTC review places the twelve word starts within 49 ms of the model estimates in this example. End estimates differ more. Whisper's short-line timestamps disagree more; its unprompted transcripts and the CTC model's imperfect greedy transcript are retained. Neither check establishes universal accuracy or exact acoustic boundaries.
CPU reproduction · Voice and script round trip · Export evidence · CTC timing review · Whisper review · Reopen the project
Footage: Michael Burrows / Pexels, under the Pexels License. Original links and checksums are included. Fictional coffee concept; no affiliation or endorsement. Generic Kokoro narration. Manrope and DM Sans (OFL).