The README shows three static screenshots. ClipForge is a video tool, and the things that make it worth using are all motion: karaoke caption timing, auto zoom punch-ins, and the 9:16 crop cutting between speakers like a camera switch. None of that is visible in a still image.
What would help most
A short before-and-after: a 16:9 source clip on the left, the exported 9:16 captioned version on the right. Even 15 seconds would do.
Why we can't just generate one
scripts/seed-demo.ts builds its demo video from ffmpeg's testsrc2 synthetic test pattern, because it exists to exercise the UI offline without API calls. Capturing that would show the app working on colour bars, which looks worse than no demo at all. A real demo needs real footage with a real person talking, so speaker tracking and captions have something to work with.
How you can help
- If you use ClipForge and are happy for a clip to appear in the README, please link one here. Something you own the rights to, ideally a talking-head video where auto-reframe and captions do visible work.
- Alternatively, a CC-licensed or public-domain talking-head clip we could use, with the licence noted.
Credit in the README for anyone who contributes footage. If you would rather build the capture tooling than supply footage, say so here and we can split it.
The README shows three static screenshots. ClipForge is a video tool, and the things that make it worth using are all motion: karaoke caption timing, auto zoom punch-ins, and the 9:16 crop cutting between speakers like a camera switch. None of that is visible in a still image.
What would help most
A short before-and-after: a 16:9 source clip on the left, the exported 9:16 captioned version on the right. Even 15 seconds would do.
Why we can't just generate one
scripts/seed-demo.tsbuilds its demo video from ffmpeg'stestsrc2synthetic test pattern, because it exists to exercise the UI offline without API calls. Capturing that would show the app working on colour bars, which looks worse than no demo at all. A real demo needs real footage with a real person talking, so speaker tracking and captions have something to work with.How you can help
Credit in the README for anyone who contributes footage. If you would rather build the capture tooling than supply footage, say so here and we can split it.