Sharing this since a few AI video tools have come up here recently.

It pulls the spoken script out of a TikTok with the exact second each line lands, and keeps on-screen captions separate from the dialogue. That separation matters more than it sounds, because the line that actually sold a video is often three words burned into a corner and never spoken out loud.

It also strips burned-in captions before a re-cut, and breaks a reference clip into numbered shots with start and end times, so you can see how it opens, demonstrates and closes.

Mostly useful if you work with TikTok, Reels or Shorts and keep rewatching the same clip trying to remember what happened at second four.