Kling 3 Motion Control
Kling 3 Motion Control transfers the movement of a reference video onto the character in a reference image, and it is best when you own a clip whose motion you want a different person to perform.
video · model kling-3-motion · catalog model-catalog@2026-10-06.v5 · pricing customer-pricing@2026-11-05.v15.Endpoint
POST https://api.roasos.com/v1/video/kling-3-motionSend your API key as a bearer token. The key needs the scope api:write. Every paid request also needs an Idempotency-Key header.
The envelope fields quote_only, quote_id, webhook_url and wait_timeout_seconds behave the same on each model endpoint. Refer to Jobs and Quotes and credits.
Good at
The reference image supplies the character, the background and the other visible elements. The reference video supplies the body motion and, with character_orientation: "video", the camera behaviour. Use it for dance, gesture and presenter moves that must match a source clip exactly. The standard tier is the cost-effective mode for portraits and simple animations; the pro tier is the choice for full-body motion and detailed clothing. Write the prompt about the subject, the wardrobe, the expression and the style, not about the camera, because the reference video already owns the camera. Keep the reference video short: it decides both the output length and the price. It is the only video model on roasOS that takes a reference video as its motion source, and the only one that requires a recorded media-rights attestation. It does not accept a text-only or image-only request, a last frame, or audio input.
Pricing
The unit is one clip. Pricing keys: kling-3-motion-pro, kling-3-motion-standard.
The catalog has no credit example for this model. Send quote_only: true to get the price.
Modes
motion-control
Provider endpoints: fal-ai/kling-video/v3/pro/motion-control · fal-ai/kling-video/v3/standard/motion-control.
Request fields
| Field | Type | Required | Default | Rule |
|---|---|---|---|---|
character_orientation | enum | yes | image | image, video |
extend_of_generation_id | string | no | — | UUID |
image_url | string | no | — | https URL; ≤ 4,096 characters |
keep_original_sound | boolean | yes | — | — |
media_rights | object | yes | — | { accepted, statement_version } |
prompt | string | no | — | ≥ 1 character; ≤ 2,500 characters |
quote_id | string | no | — | UUID |
quote_only | boolean | no | false | — |
scene_id | string | no | — | UUID |
tier | enum | no | standard | standard, pro |
video_url | string | yes | — | https URL; ≤ 4,096 characters |
wait_timeout_seconds | integer | no | 0 | 0–60 |
webhook_url | string | no | — | https URL; ≤ 4,096 characters |
The schema is strict. A field that is not in this table gets 400 unsupported_option. The field names are the roasOS role names. One role has one name in every family.
Media
This endpoint accepts these media roles:
image_url— first frame (or the single source image)scene_id— first frame — a scene this workspace ownsvideo_url— the source video
| Role | Items | Maximum size | Content types | Duration |
|---|---|---|---|---|
image_url | ≤ 1 | 10 MB | image/jpeg, image/png, image/webp | — |
video_url | ≤ 1 | 100 MB | video/mp4, video/quicktime | — |
Every other role is refused: avatar_id, elements, product_id, product_images, references, reference_images, end_image_url, scene_image, video_urls, audio_urls, audio_url, voice_id. A refused role gets 400 unsupported_option.
https://cdn.roasos.com that your workspace owns. Any other URL gets 400 reference_media_not_workspace_owned. Refer to Reference media and uploads to learn how to put a file there.Prompt
- Limit: 2,500 characters. No roasOS wrapper; the prompt goes verbatim, or the field is omitted when empty.
- No tags:
@Element1exists at fal only withelements, which roasOS does not expose. - Craft: name the subject and the wardrobe ("woman in a red linen dress, hair tied back"), the expression ("calm, small smile"), and the style ("natural daylight, realistic skin"). Do not describe camera moves or the motion itself; the video owns both. Keep it short: the image and the video carry the information.
Not exposed
These catalog entries describe omitted inputs, pinned values and source constraints. They do not imply that the provider accepts every listed field.
duration_seconds— pinned tomeasured from video_url.aspect_ratio— not accepted by this public endpoint.resolution— pinned tonamed by tier.generate_audio— not accepted by this public endpoint.negative_prompt— not accepted by this public endpoint.seed— not accepted by this public endpoint.references— not accepted by this public endpoint.elements— not accepted by this public endpoint.avatar_id— not accepted by this public endpoint.product_id— not accepted by this public endpoint.end_image_url— not accepted by this public endpoint.
Result
The result kind is video. Always present: url.
{
"data": {
"id": "gen_…",
"kind": "video",
"model": "kling-3-motion",
"status": "completed",
"result": {
"url": "https://cdn.roasos.com/workspaces/ws_01/result"
}
}
}The files are immutable objects on https://cdn.roasos.com. A result URL is a valid media input for your next request.
Example
The three steps are always the same: quote, run, poll.
# 1. Quote. Nothing is spent.
curl -X POST https://api.roasos.com/v1/video/kling-3-motion \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-kling-3-motion-0001" \
-H "Content-Type: application/json" \
-d '{"quote_only":true,"character_orientation":"image","image_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/image_url.png","keep_original_sound":false,"media_rights":{"accepted":true,"statement_version":1},"video_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/video_url.mp4"}'
# 200 { "data": { "kind": "quote", "quote_id": "q_7Hk...", "credits": ..., "expires_at": "...",
# "confirmation": { "model": "kling-3-motion" } } }
# 2. Run. Same body, same Idempotency-Key, quote_id in place of quote_only.
curl -X POST https://api.roasos.com/v1/video/kling-3-motion \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-kling-3-motion-0001" \
-H "Content-Type: application/json" \
-d '{"quote_id":"q_7Hk...","character_orientation":"image","image_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/image_url.png","keep_original_sound":false,"media_rights":{"accepted":true,"statement_version":1},"video_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/video_url.mp4"}'
# 201 { "data": { "id": "gen_...", "status": "queued", "credits_reserved": ... } }
# 3. Poll, then read the result.
curl https://api.roasos.com/v1/jobs/gen_.../status -H "Authorization: Bearer roasos_sk_live_..."
curl https://api.roasos.com/v1/jobs/gen_.../result -H "Authorization: Bearer roasos_sk_live_..."The deprecated path
POST /v1/generations still reaches this model for one minor version. It is deprecated. This is how it maps:
Not available on the public API yet. POST /v1/generations returns 503 public_model_unavailable for model: "kling-3-motion", and the alias will stay refused because it cannot carry media_rights. The app reaches the model through its internal UGC route (image, referenceVideo, characterOrientation, keepOriginalSound, quality, rightsAccepted, rightsStatementVersion, rightsPolicyVersion); at v1.1 these become image_url, video_url, character_orientation, keep_original_sound, tier, media_rights{}. The agent does not offer this model.
See also
llms.txtfor this model — the same facts as one plain-text file for an agent.GET https://api.roasos.com/v1/catalog/kling-3-motion— the machine-readable form: enums, defaults, limits and prices. Refer to the API reference.- Jobs — how to poll a job and read its result.
- Provider endpoints:
fal-ai/kling-video/v3/pro/motion-control·fal-ai/kling-video/v3/standard/motion-control.
MiniMax H3 Styles
MiniMax H3 Styles makes video with sound in one of five fixed looks: VHS tape, 1970s hand-painted toon, low-poly 3D, hand-drawn animation or 16-bit pixel art. It renders at 768p, in any whole length from 5 to 15 seconds.
Kling 4.0
Kling 4.0 makes video from text, from one first frame, or from reference media. It renders at 720p or 1080p, in any whole length from 3 to 30 seconds. It is the Kling model for new video: it keeps a person or a product the same from shot to shot.