Are these YouTube-style chapters?
No. The tool creates smaller line/visual beats based on the speaker's actual sentence endings and pauses. There is no fixed 1.8–3.6 second cutoff, so a sentence can run longer when that is what the speech requires.
Why are some lines 2 seconds and others 5 seconds?
The grouping follows word timing, sentence punctuation and real pauses. Duration is a soft visual preference, not a hard cutoff, so complete thoughts stay together.
Can I paste an existing YouTube/Instagram transcript?
Yes if it includes usable timestamps. If it is plain untimed text, upload the matching media to create real timing instead of fabricated timing.
Can I use it for a three-hour narration?
Yes. Long media is chunked internally so the final timeline keeps absolute timestamps.