Create Video

Endpoint

POST /v1/videos

Returns HTTP 202 immediately. Poll GET /v1/videos/{id} for the result.

Request

string

required

Bearer <your-token>

string

required

application/json

string

Any unique string. Reusing the same key returns the existing job instead of creating a new one. See Idempotency.

Body

string

required

Model ID from GET /v1/models.

string

required

Text description of the video to generate. Max 4000 characters.

string

Elements to avoid in the output. Max 4000 characters.

string

default:"16:9"

Output aspect ratio. Common values: 16:9, 9:16, 1:1, 4:3, 3:4, 21:9. Seedance also supports adaptive.

Veo 3.1 models ignore this field — output is always 16:9. Omni honors 16:9/9:16 correctly.

number

default:"5"

Clip length in seconds. Range depends on the model:

Kling v3.0, Kling Video O3 — 3 or 15 only (no other value accepted)
Kling v2.x, Kling Turbo, Kling Video O1 — 5 or 10 only
Seedance 2.5 — 4 to 30 (any integer), or -1 for auto-length
Other Seedance — 5 to 10 (any integer)
Omni — 3 to 10 (any integer)
Wan 3.0 — 2 to 30 (any integer), or -1 for smart/auto-length
MiniMax H3 — 4 to 15 (any integer)
Veo 3.1 — fixed at 8, this field is ignored

Check capabilities.seconds in GET /v1/models for the exact list per model.

string

Output resolution. Supported values: 480p, 720p, 1080p, 4k. Availability depends on model — Kling v3.0-4k and Kling Video O3-4k default to 4K. Veo 3.1 and Omni support 720p (default) and 1080p.

number

default:"0.5"

Prompt adherence for Kling v1.x / v2.x models. Range 0–1. Higher = more literal.

number

default:"0.5"

Prompt adherence for Kling v3.0+ and Kling Video O3 models. Range 0–1.

integer

Reproducibility seed. Pass the same seed to regenerate an identical result.

string

Start frame image URL. Required for most image-to-video models (Kling v1–v3, Seedance, omni/image-to-video). Not used by Kling Video O3 — for O3, pass reference images via images[] instead.Don’t have a public URL? Upload the file with Upload Media first and pass the returned url here.

string

End frame image URL (start→end interpolation). Supported by Seedance and Kling v3.0/O3 models.

string

End frame image URL. Alias for end_image — accepted by Kling v2.5-turbo and omni/image-to-video. For omni, the model interpolates between image (start) and last_image (end); both are optional but at least image or images[0] must be set.

string

Reference video URL. Used for motion-control, video-edit, and reference-to-video models. Upload a local file first via Upload Media if you don’t have a public URL.

string

Audio URL for lipsync models (e.g. kling/lipsync/audio-to-video). Provide a public URL to the audio file. Use together with video (face video).

array

Array of image URLs. Behavior depends on the model:

Kling Video O3, omni/reference-to-video — primary reference images for reference-to-video generation. image is NOT required; pass all reference frames here (omni/reference-to-video accepts 1–10). The model derives style/character from them rather than treating any one of them as a literal start/end frame. URLs must be publicly accessible from international servers — Shopee (susercontent.com), Google Drive, and other CDNs that geo-restrict or require auth will fail with invalid_parameters. Upload via Upload Media first if you’re not sure.
omni/image-to-video, wan3/image-to-video, minimax-h3/image-to-video — images[0] acts as a frame alias: no image set → images[0] becomes the start frame; image already set → images[0] becomes the end frame. Unlike the reference-to-video models above, these only ever accept 2 frames total (start + end) — not a flexible reference set. Sending more than 2 images total (across image + images[] combined) is rejected with invalid_parameters rather than silently dropped.
Other models — this field is ignored.

boolean

default:"true"

Generate audio synchronized with the video. Supported by Seedance 2.5.

boolean

Generate audio for Kling v3.0+, Kling Video O3, Kling v2.6 models.

string

default:"customize"

Camera shot type for Kling v3.0+ and Kling Video O3. Values: customize, close_up, medium_shot, full_shot, over_shoulder, aerial.

array

Multi-segment prompts for Kling v3.0+ and Kling Video O3. Each item: { "prompt": "...", "duration": 3 }. The sum of all duration values must equal the top-level duration.

boolean

Lock camera movement throughout the clip. Seedance only. Pass true to prevent camera panning or zooming.

string

Kling element ID for a persistent character or product across multiple videos.

string

Voice ID for Kling lipsync and avatar models (e.g. genshin_klee2).

object

Watermark configuration for Kling models. Example: { "enabled": false }.

string

Your internal task ID. Stored alongside the job and returned in all responses — useful for reconciling logs between your system and ours.

Examples

Text to video (Kling v3.0, 15s):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kling/v3.0-pro/text-to-video",
    "prompt": "A wooden boat drifting across a calm bay at dawn",
    "aspect_ratio": "16:9",
    "duration": 15,
    "cfg_scale": 0.5
  }'

Image to video (Seedance 2.5, 30s with audio):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance/2.5/image-to-video",
    "prompt": "The subject walks forward slowly",
    "image": "https://example.com/photo.jpg",
    "duration": 30,
    "generate_audio": true,
    "resolution": "1080p"
  }'

Start–end frame (Kling v3.0):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kling/v3.0-pro/image-to-video",
    "prompt": "Camera pans slowly from left to right",
    "image": "https://example.com/first-frame.jpg",
    "end_image": "https://example.com/last-frame.jpg",
    "duration": 3
  }'

Text to video (Omni, 6s, portrait, 1080p):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "omni/text-to-video",
    "prompt": "A red bicycle leaning against a brick wall, soft morning light",
    "duration": 6,
    "aspect_ratio": "9:16",
    "resolution": "1080p"
  }'

Start + end frame to video (Omni):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "omni/image-to-video",
    "prompt": "Smooth camera transition between the two scenes",
    "image": "https://images.pexels.com/photos/21050390/pexels-photo-21050390.jpeg",
    "last_image": "https://images.pexels.com/photos/12751943/pexels-photo-12751943.jpeg",
    "duration": 5
  }'

Text to video (Veo 3.1, 1080p — duration/aspect_ratio have no effect on this model):

curl -X POST https://hypergate.store/v1/videos \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "veo3-quality/text-to-video",
    "prompt": "A red bicycle leaning against a brick wall, soft morning light",
    "resolution": "1080p"
  }'

Response `202`

{
  "object": "video",
  "id": "3fa85f64-5717-4562-b3fc-2c963f66afa6",
  "model": "kling/v3.0-pro/text-to-video",
  "status": "in_progress",
  "created_at": 1790000000,
  "external_task_id": "my-internal-job-456"
}

Getting Started

API Reference

Guides

Endpoint

Request

Body

Examples

Response `202`

​Endpoint

​Request

​Body

​Examples

​Response 202

Endpoint

Request

Body

Examples

Response `202`