Sume motion routes by length: Kling, Recast and Genjutsu
Three Sume routes drive a result from a source clip, each with a length rule: Kling takes duration_seconds, Recast follows the source, Genjutsu a range.

Sume offers three ways to drive a result from a source clip, and each one handles length differently. Kling 3.0 Motion Control needs duration_seconds from 1 to 30 and its output follows the driving video. H3 Max Recast takes 5 to 30 seconds and uses the source video's length rounded up. Higgsfield Genjutsu, listed only when its provider is configured, accepts 4 to 30 seconds. Know which rule applies before you send a long clip.
Prices matter here because all three bill per second of output.
The three length rules
Ranges and resolutions are from the catalog and the Kling route; confirm the live rows.
| Route | Inputs | Length rule | Resolutions |
|---|---|---|---|
Kling 3.0 Motion Control (POST /v1/kling/3.0/motion-control) | a still or avatar plus motion_video_url | duration_seconds 1 to 30, required; output follows the driving video | no resolution field |
h3-max-recast | a source video plus 1 to 4 person photos | 5 to 30 s; duration is the inspected source length, rounded up | 768p, 1080p |
higgsfield-genjutsu | a source video plus 1 to 8 reference images | 4 to 30 s | 480p, 720p |
What Kling costs by length
Kling Motion Control's list is $0.126 per second in Sume's pricing table, so with the 1.25 margin it is about $0.1575 per second. A 5-second clip is about $0.79 and a 30-second clip about $4.73, rounded up to the cent. There is no resolution knob: the standard route has none, and output length follows the driving video.
- 5 s: about $0.79.
- 15 s: about $2.37.
- 30 s: about $4.73.
A Kling request
Fields are exactly those the schema accepts: image_url, motion_video_url, duration_seconds, and optional prompt, keep_original_sound and character_orientation. Both URLs must be public HTTPS addresses.
curl -X POST https://api.sume.com/v1/kling/3.0/motion-control \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: kling-mc-len-001" \
-d '{
"image_url": "https://example.com/character.png",
"motion_video_url": "https://example.com/dance-8s.mp4",
"duration_seconds": 8,
"character_orientation": "video"
}'How to pick
If the photo should perform someone else's motion, use Motion Control. If people in an existing clip should be replaced, use Recast. If neither fits because you need a different framing, generate a new clip on a reference row. Avoid padding: a 12.4-second Recast source bills as 13 seconds. Read the Video Router page, the MCP tools page for kling-motion-control_create, and the API reference.
Sources
Related posts
More in Models
- Nano Banana reference image limits: Lite, 2 and Pro compared
Nano Banana 2 Lite takes up to 14 object images, Nano Banana 2 takes 10 object, 4 character and 3 style, Pro takes 6 object and 5 character.
- Pixal3D multi-view: prepare the input images
Pixal3D added multi-view inference in September 2026 under an MIT license. Sume has no 3D, but reference-image edit can prepare input views.
- PixVerse V6 native audio and camera work: what to check in an API
PixVerse's blog lists V6 with camera work and native audio, plus R2 and a $439M Series C total. How to test those claims against any video API's catalog.
- PixVerse V6 adds native audio: which Sume video ids make sound
PixVerse V6 ships camera work and native audio. The Sume video docs name no PixVerse id, so here is how to find models that generate audio and read the flag.
Written by Sume