H3 Max Recast: 1-4 person photos, 768p default, 1080p opt-in
H3 Max Recast swaps the people in one 5-30 s video using 1-4 reference photos, one per new person. Output is 768p by default; 1080p is opt-in.

h3-max-recast replaces the primary people in one source video with the people in 1-4 reference photos. Send one photo for each new person. The output is 768p by default, and 1080p is opt-in. The source video must be 5-30 seconds, with no single shot longer than 15 seconds.
Limits
The limits come from the Video Router catalog row.
| Item | Value |
|---|---|
| Source video | video_url, 5-30 s, no shot over 15 s |
| Reference photos | reference_image_urls, 1-4, one for each new person |
| Prompt | Optional, up to 2000 characters |
| Resolution | 768p (default) or 1080p |
| Duration billed | The length of the source, rounded up |
| Text-only generation | Not supported |
What stays the same
The output keeps the source motion, camera, cuts and sound. You are changing who is in the shot, not how it moves. Recast is selected only explicitly: sume/auto never routes to it.
Price
The list rate is $0.30 per second at 768p and $0.45 at 1080p (fal 2026-10-01). Times 1.25 that is $0.375 and $0.5625 per second. A 10 s source at 768p is 10 x $0.375 = $3.75, and at 1080p 10 x $0.5625 = $5.625 before cent rounding. A 30 s source at 768p is $11.25.
Example
Request on the Video Router. video_url is accepted only by this row, higgsfield-genjutsu and gemini-omni-flash-1.1.
{
"model": "h3-max-recast",
"video_url": "https://example.com/source-12s.mp4",
"reference_image_urls": [
"https://example.com/person-a.jpg",
"https://example.com/person-b.jpg"
],
"resolution": "768p"
}Sources
Related posts
More in Media tools
- Hailuo 3.0 clips in a Sume timeline: set output.fps or omit it
Cutting 24 fps AI clips into a timeline at the wrong frame rate causes judder. Sume's Timeline output.fps accepts 24, 25, 30 or 60, or matches your sources.
- Hand an AI clip to an editor: MP4, PNG frames and WAV on Sume
What Sume can deliver for a finishing editor: MP4 clips, lossless PNG stills at source size, and sample-exact WAV audio. No EXR or HDR is documented.
- Headphones listing video: spec card on the left, clip on the right
Build a 16:9 listing video for headphones with Timeline compose: a vertical stack with the spec card left, 0.35 ratio, clip on the right, $0.02 a shot on Sume.
- Hotel lobby screen music: a 2-minute welcome video loop
Make a 2-minute welcome video with a soft jazz-lounge bed for a hotel lobby screen: one generation and a two-minute render, $0.325 on Sume.
Written by Sume