How to cut a long video into short clips that end on a finished sentence
Give Vizard Agent the video and a maximum clip length. Vizard Agent transcribes it, finds candidate passages that fit inside your limit, then verifies every boundary against the word timings and adjusts any that would have cut a sentence in half, so each clip starts and ends on a complete thought.
What do you type into Vizard Agent?
Give the ceiling and the source. Vizard Agent works out where the natural boundaries are on its own, so there's no need to nominate timestamps, and a single sentence naming the limit is genuinely the whole brief for this job. Say how many clips you want, or let it choose.
Prompt
What does Vizard Agent actually do?
Cutting on the clock is trivial and useless: a twenty-second clip that stops halfway through a sentence is unpostable. So Vizard Agent chooses the boundaries from the speech, then spends half the run checking boundaries it has already chosen, which is what caught a clip that would have ended mid-sentence. In the session behind this article it:
- Downloaded the video from the link, analysed its metadata and transcribed the audio to find natural cut points.
- Found high-quality, non-overlapping candidate passages under the 20-second limit, and confirmed the segments carried word-level timings.
- Extracted exact word-aligned start and end times for all five clips, then moved the end of the second clip because it would have cut off mid-sentence.
- Verified the boundaries of every other clip in turn, printing the surrounding lines to check each one.
- Cut the clips at exact frame boundaries and extracted a frame from each to confirm the picture.
- Delivered five separate clips with their own links, one probed at 1920x1080, H.264, 30fps and exactly 20.0 seconds.
When does this not work well?
The transcript is what makes the boundaries good, so anything it doesn't capture is invisible to the selection. A clip can be complete and still make no sense if its sentence refers to the previous ten minutes. Silent passages give the transcript nothing to work with, and a scored video's music breaks at the joins.
Common questions
Will the clips cut off mid-sentence? No. Vizard Agent sets the boundaries from word timings and re-checks each one before cutting.
Can it work from a link? Yes. Vizard Agent downloads the video and works from the file.
What if the video has no speech? Then there are no spoken boundaries to find, and asking Vizard Agent to cut on shot changes instead is the better request.