Swap the lead in a finished AI episode: H3 Max Recast for 20 s
H3 Max Recast replaces the people in a 5-30 s source with 1-4 photos, keeping motion and sound. A 20 s source is $7.50 at 768p and $11.25 at 1080p on Sume.

To change who plays the lead in a finished episode, send the video and one to four photos to H3 Max Recast on Sume. It keeps the source's motion, camera, cuts and sound and swaps the people. A 20-second source costs $7.50 at 768p, the default, and $11.25 at 1080p, at the billed rate of the list price times 1.25.
What it takes
| Item | Value |
|---|---|
| Model id | h3-max-recast |
| Input | One source video plus 1 to 4 person photos, one photo per new person |
| Source length | 5 to 30 seconds, with no shot longer than 15 s |
| Prompt | Optional, up to 2000 characters |
| Resolution | 768p (default) or 1080p |
| Duration billed | The inspected source length, rounded up |
| List price per second | $0.30 at 768p, $0.45 at 1080p |
| Billed per second | $0.375 at 768p, $0.5625 at 1080p |
| Text-only generation | Not available |
The arithmetic
Billing follows the source length rounded up, so a 20-second episode costs 20 x $0.375 = $7.50 at 768p and 20 x $0.5625 = $11.25 at 1080p. A 30-second source, the maximum, is $11.25 or $16.88. Those are the prices for one pass, and a failed attempt or a second try is a second charge.
As a season line, ten episodes of 20 seconds recast at 768p would be $75.00, and the same at 1080p $112.50.
Prepare the source
The no-shot-over-15-seconds rule is the one that surprises people. If your episode has a single 25-second take, cut it into two shots first, because the model needs cuts to handle. Inspect the file with the video-inspect endpoint to see its length, and trim extra seconds off the end so you do not pay for them.
Choose the photos
- Use one clear, front-facing photo per person you want in the shot.
- Match the number of photos to the number of people being replaced, up to four.
- Use only people who have agreed to appear, and do not use a real person's likeness for a character without consent.
- Compare the first still of the result with the original before approving, since a swap changes the face but keeps everything else.
Recast or re-render
If the lead is only wrong in some shots, re-rendering those shots with a reference image is often cheaper than recasting a whole 20-second episode. Recast is the right tool when the performance, camera move and sound are already right and only the face is not. For your own avatar in a short cameo, the Face Swap route is the other option.
Sources
Related posts
More in Models
- MiniMax H3 Max on Sume: video, lip-sync and recast, which id to call
Three Sume ids carry the MiniMax H3 Max name: minimax-h3-max for video, a lip-sync route and h3-max-recast. What each takes, its window and its rate.
- Two reference images, one clip: Omni's cat-and-yarn pattern on Sume
Google's docs show two images, a cat and yarn, producing one video. Here is the Sume request with IMAGE_REF tokens and what changes with a brand product.
- Video edit prompt: say what stays, then what changes (Omni Flash)
A prompt pattern for Gemini Omni Flash 1.1 video edit on Sume: one change per pass, an explicit keep clause, and what the video_url route fixes for you.
- Does Wan 3.0 do 4K? Claim check against vendor pages and Sume
Alibaba lists 480p, 720p and 1080p for Wan 3.0, Morphic says no 4K, and Sume rejects 4k for wan-3.0. Where 4K exists on Sume and at what clip length.
Written by Sume