AI video models with no aspect_ratio option: Recast, Genjutsu, Grok
Three Sume video rows reject aspect_ratio: H3 Max Recast, Higgsfield Genjutsu and Grok Imagine Video 1.5. What sets the output frame instead, and what to send.
Three rows in Sume's video catalog advertise an empty list of aspect ratios, which means a request that includes aspect_ratio is refused. They are H3 Max Recast, Higgsfield Genjutsu and Grok Imagine Video 1.5. In all three, the frame comes from an input you supply, so asking for a ratio would contradict it.
The Gemini Omni Flash 1.1 edit path has the same rule on the Video Router: aspect_ratio is rejected when you send a source video to edit.
What decides the frame in each row?
The table lists the row, what sets its frame, and what to omit.
| Row | Frame comes from | Also rejected |
|---|---|---|
h3-max-recast | The source video; output keeps motion, camera and cuts | frame_images, generate_audio, seed, audio references |
higgsfield-genjutsu | The source video; output preserves length and framing | frame_images, generate_audio, audio references |
grok-imagine-video-1.5 | The input image | generate_audio, bitrate_mode, end frame, references |
What if I need a different ratio?
Change the input, not the request. For Grok, crop or pad the source image to the shape you want before upload. For Recast and Genjutsu, reframe the source video first, since the output keeps its framing. If you cannot change the input, move to a row that lists the ratio, such as Wan 3.0 or MiniMax H3, but note those do not do person swaps.
This is the same catalog-driven rule behind the other 400s: Sume refuses a value a row does not advertise rather than dropping it. See the unsupported parameter post for size, seed and provider.options.
How do I check a row before sending?
Read supported_aspect_ratios from GET /v1/videos/models. An empty list means do not send the field. A small guard in your client that omits aspect_ratio whenever that list is empty prevents the whole class of errors.
Sources
Related posts
More in Developers
- AI video person swap rejects long takes: find shots over 15 s
H3 Max Recast on Sume needs 5 to 30 seconds with no single shot over 15. Find the long takes with ffmpeg scene detection, then cut them before you pay.
- AI voice reads Spanish with an English accent: two causes to check
Spanish read with an English accent usually means the language field was omitted or the voice is tagged for another language. How Sume's TTS 409 works.
- Amazon Sponsored Brands video: Sume fps 24-30 fit; 60 and 4K do not
Amazon lists 23.976 to 30 fps and 1280x720, 1920x1080 or 3840x2160. Sume trim offers 24, 25, 30 or 60 fps and stops at 2160 per edge. Which settings line up.
- Animate a picture with Kling 3 via API: Python first-frame request
To animate one picture with Kling 3 on Sume, send it as a first_frame in frame_images to POST /v1/videos with model kling-3. Full Python script included.
Written by Sume