AI video generation API: Kling, Veo, Seedance and Hailuo

One endpoint starts a video job on Kling 3.0, Seedance 2, Hailuo, Wan and other models, from a prompt or a still image. Veo 3.1 is available through the raw passthrough. Jobs run as tasks and the charge is the model's actual cost for that clip.

Who it is for

Studios and creators producing short clips and ads
Apps that animate user images
Teams testing several video models before committing to one

How it works

POST a prompt (and image_urls for image-to-video) with model-specific options in input.
A hold covering the expected cost is placed on your credit while the clip renders.
When the task finishes, output.videos has the URL and the hold is replaced by the actual charge.

Endpoints

Method	Path	What it does
`POST`	`/v1/videos/generations`	Start a video task
`POST`	`/v1/generations`	Other generation models (lip-sync, upscaling)
`GET`	`/v1/tasks/{id}`	Poll for the video URL

Models

Pass one of these as model (or voice_id for voices). The full, current list with prices is at /v1/models.

kie/kling-3.0/video
kie/bytedance/seedance-2
kie/bytedance/seedance-1.5-pro
kie/grok-imagine/image-to-video
kie/hailuo/02-text-to-video-pro

Example

curl https://omni-api.rcinc.app/v1/videos/generations \
  -H "Authorization: Bearer $OMNI_KEY" -H "content-type: application/json" \
  -d '{"model":"kie/kling-3.0/video","prompt":"Slow push-in on a rainy Tokyo alley at dusk","input":{"duration":"5"}}'

Pricing

Charged per clip or per second, depending on the model. Veo 3.1 Fast at 1080p costs about $0.37 per clip on Omni against Google's $1.20; Hailuo 2.3 Standard costs about $0.17 against $0.28. See the pricing page for the full comparison.

Limits

Video jobs take from under a minute to several minutes.
Each model has its own options (duration, resolution, audio), passed in the input object.
Premium or long clips may need max_cost_usd set so the hold covers them.