Side project · AI automation

One long video in. Ready-to-post clips out.

Turning a podcast or stream into short clips usually means hours of scrubbing, cutting and captioning. Clipwise does the whole job: it finds the moments worth sharing, cuts them for every platform and adds the captions.

Try it yourself1Press Run2Watch each step finish
✓Transcribed word by word
✓Found the best moments
✓Cut tall and wide versions
•Adding captions
Ready to post
Project
My own side project
What I did
Pipeline, AI, editor, web app
Built with
Python, Whisper, ffmpeg, AI
Status
Live, in daily use
Try it yourself1Tap a peak2See why it was picked
Most replayed parts of a 58-minute podcastfrom YouTube
Clip 1 of 3 · 12:40 to 13:25 "I lost the whole company in a week." The most replayed part of the episode, and it opens with a hook in the first second.
94 clip score
How it picks moments

It trusts what viewers rewatched before what the AI thinks.

Guessing good moments from a transcript is hard. Real viewers already voted with their replays, so that comes first. The AI then finds where each moment starts and ends cleanly. If the AI is unavailable, a simple scoring rule still does a decent job, so it never stops working.

  • 1Start from what viewers rewatched (YouTube's Most Replayed data)
  • 2AI finds a clean start and end around each moment
  • 3No AI available? A scoring rule rewards hooks and pace, and skips the ums
Why it matters for your business

Most businesses have a job like this.

Something repetitive that eats hours every week. Clipwise is the same kind of AI automation I build for clients, just pointed at video.

  • Runs on its own once set up
  • AI where it helps, simple rules where they are enough
  • Keeps working if a service is down
  • Hours of work become minutes
Next projectCool Contours
Your projectGot a job AI could do? Get a free quote