YouTube already knows. Every video carries a replay graph built from millions of people dragging the scrubber backwards. Rewound reads that graph and cuts where the peaks are. Paste a link, get three clips back. Nothing to install.
Counted in clips you keep, not hours you upload. A four-hour podcast costs the same as a four-minute one.
Rewound pulls the replay data before it downloads a single frame of video, so it knows where to cut before it starts. Everything runs on our machine.
The replay heatmap arrives with the video’s metadata. We smooth it, then throw away the opening spike, because every video’s first seconds are its most replayed and never the interesting part.
The highest non-overlapping moments, weighted toward the setup that precedes each payoff. In and out points snap to sentence boundaries, so nothing starts mid-word.
Reframed to 9:16, captions burned in word by word, the source credited on screen. A title, description and hashtags come with every clip.
A language model reads the transcript and estimates which passage sounds compelling. It is a guess about text, made without ever seeing how a single viewer behaved.
Reads the record of what viewers actually did. Not a prediction of what should have been interesting. Evidence of what was.
Paste the video you would clip first. Clips come back cut, captioned and ready to post.
No account, no card. Three clips, free.