the short version
- ChatCut’s remake keeps a clip’s motion, camera moves and shot timing, and draws it again with new people, products, clothes, a new place or a new look. You get back a new video clip.
- HyperFrames’ Reverse watches a video and rebuilds every layer (text, graphics, layout and timing) as a HyperFrames project. You get back a project you can change line by line and render again.
- I use ChatCut and HyperFrames for my videos, but I haven’t used either of these two features on a real project yet. This page comes from each tool’s own description, checked on 5 Nov 2026.
side by side
| ChatCut remake | HyperFrames Reverse | |
|---|---|---|
| what you get | A new video clip in your ChatCut project | An editable HyperFrames project (scenes written as web pages), and a render of it |
| what changes | The people, products, clothes, place or look you name. Reference images set the new look | Nothing at first: it rebuilds the video as it is. Then you change the words, colours, branding or data |
| length | 4–30 seconds of the source, in whole seconds | The whole video. A rebuild takes about 20–50 minutes |
| people on screen | Recast from your photos | Re-shot with HeyGen, following the original’s body and camera movement, with the original voice moving the lips |
| text on screen | Object-swap keeps it. Motion-transfer turns large text into garbled letters | Rebuilt as real text you can edit |
| sound | Keeps the source audio | Keeps the original soundtrack |
| checking | You watch the result | Compares its render with the original frame by frame, and lists missed cuts and the worst moments |
| cost | ChatCut credits. 720p by default; 1080p costs about 2.5 times as much | Needs a paid HyperFrames Pro plan |
| where it runs | Generated in ChatCut’s cloud with Seedance 2.5. Motion-transfer reads the motion on your own computer first | Built on your computer by an AI agent with the HyperFrames skills. Renders in HyperFrames’ cloud by default, or on your computer. HeyGen makes the re-shot people |
the two remake modes
- object-swap
- Changes only the people, products or clothes you name. Everything else stays, including other people and text on screen. It can keep a little of the original face.
- motion-transfer
- Draws the whole frame again in a new place or style, keeping the motion and camera. Recast people match their photos better, but large text burned into the video comes out garbled.
use this when
- ChatCut
- You want to change who or what is in the footage. Recast a role play, put your own product in, or restyle a scene.
- HyperFrames
- You want an editable copy of a video’s design and motion. Remake an explainer you like with your own words, colours, branding or data, or get back a project for an old video whose source files are lost.
- neither
- The video belongs to someone else and you don’t have the rights to it.
things I’d check
- Your rights to the video. Use both on your own videos, or ones you have the rights to. Closely rebuilding someone else’s video can be copying.
- The cost before it starts. In ChatCut’s own chat, a card shows the settings and the cost before anything runs. If an AI agent starts the remake from outside the app, the credits go at once, so ask it for the length and cost first.
- Very different shapes. Both ChatCut modes keep the source’s poses, outlines and sizes. Swap a small bottle for a big box and it can look wrong.
- The people in the video. Reverse re-shoots everyone on screen with HeyGen, from the original clip and voice. Check the people in it are happy with that, or ask for them to be cut from the original instead.
Based on ChatCut’s and HyperFrames’ own tool descriptions, checked 5 Nov 2026. Both change often, so check before relying on this. Related: ChatCut or HyperFrames Studio? and HyperFrames two ways.
Back to the toolkit, or see what I’ve been trying.