Video APISoon
Send a YouTube link with an API key, poll the project as JSON and download clips and caption files from signed URLs.
Get notifiedAI reframe
ReframeAny turns a recording into 9:16, 1:1, 4:5 or 16:9 clips with the crop following the person who is speaking, and centered when there is no face. You pick the format per project; custom ratios come after the beta.
Invite-only beta, free: 60 minutes a month.
Example creators and quotes. The names, numbers and teams shown are fictional until real customer stories exist.
Built for talking footage
Interviews, podcasts, talks and vlogs: with several people in view, the crop follows the person who is speaking, estimated from mouth movement, and holds each choice so it does not flicker. With one face it follows that face; footage without faces, such as gameplay or screen recordings, gets a centered crop.
Turnaround
Reframing and rendering run in the background as part of each clip. When a clip is ready you get a preview player and an MP4 download, 1080 by 1920 for 9:16, or 4K export when your source is 4K.
Object tracking · Coming after the beta
Today the crop follows faces and picks the person who is speaking. Tracking bodies, balls and on-screen text comes after the beta.
Split screen
Choose Split screen, or leave the layout on Auto, and a 9:16 clip with two people in view for most of it shows them stacked, each crop following its own face, with the captions on the seam. Split screen is 9:16 only; other formats get one crop. A screen share above the camera or blurred bars come after the beta.
Manual reframe
Open a clip's framing editor, drag the crop over the source video and set keyframes; the clip re-renders along your path without using credits. Projects made before the editor existed, and linked videos left untouched for 7 days, have no browser copy of the source to edit on. Switching single scenes between automatic and manual tracking comes after the beta.
Each feature page shows one tool. They all run on the same project, so a clip made here is ready for captions, scheduling and analytics without an export in between.
Finds the strongest moments in any long video.
Word-timed captions in three styles.
Any video to vertical, square or wide, with the crop following the person speaking.
Stock cutaways from Pexels at moments the AI picks, once the operator has set up Pexels.
Fix a misheard caption word and trim a clip by its words.
Schedule clips to YouTube Shorts and TikTok once the operator has set them up. More platforms after the beta.
Coming after the beta: see which clips bring views back to the long video.
Share projects with a team: members watch and download the finished clips. Roles and team brand kits come after the beta.
Start projects without the dashboard: call the API from your pipeline or hand it to an agent.
Send a YouTube link with an API key, poll the project as JSON and download clips and caption files from signed URLs.
Get notifiedGive an agent tools to create projects, check their status and fetch clip download links inside its own workflow.
Get notifiedFAQ
Paste a video and get it back in the format you post.
Invite-only beta, free