Wan 3.0 API
Generate up to 30 seconds of 1080p video with sound in one take, from text, an image, first and last frames, or reference images, through the imageat video API.
Live playground
Try Wan 3.0 in your browser
Run a real video generation using your imageat account, then preview and download the completed output without leaving this page.
Output
Example outputTransparent pricing
Pay only for generated seconds
No model subscription or infrastructure fee. API requests are billed in USD; generations made in the browser playground use imageat credits.
480p
Drafts and previews
$0.10 / sec
5-second video
$0.50 · 10 cr
30-second video
$3.00 · 60 cr
Browser playground: 2 credits per generated second.
720p
Social and web · default
$0.20 / sec
5-second video
$1.00 · 20 cr
30-second video
$6.00 · 120 cr
Browser playground: 4 credits per generated second.
1080p
Final delivery
$0.40 / sec
5-second video
$2.00 · 35 cr
30-second video
$12.00 · 210 cr
Browser playground: 7 credits per generated second.
Every mode uses the same rate, and audio does not change the price. Reference video and audio inputs are not supported. Failed generations are not charged.
Four ways to direct a shot
Use only a prompt, add imageUrl, set firstFrameUrl and lastFrameUrl, or pass up to 10 referenceImageUrls.
Asynchronous by design
Receive a taskId immediately and poll until the final CDN video URL is ready.
One imageat key
Use the same Bearer key, credits, run history, and response contract as other imageat models.
Quickstart
Your first request
Keep the API key on your server. Submit JSON, save the returned task and generation IDs, then poll the status endpoint.
POST /api/v1/videos
GET /api/v1/videos/status?taskId=…
curl -X POST "https://imageat.com/api/v1/videos" \
-H "Authorization: Bearer $IMAGEAT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan3",
"prompt": "A barista pours latte art in slow motion, the soft hiss of the steam wand",
"duration": "8s",
"resolution": "1080p",
"aspectRatio": "16:9",
"generateAudio": true
}'Request parameters
| Parameter | Required | Values | Description |
|---|---|---|---|
| model | Yes | wan3 | Selects Wan 3.0. |
| prompt | Yes | string | Describes subject, action, camera, lighting and sound. Put spoken lines in quotes. |
| duration | No | 2s–30s | Defaults to 8 seconds. Rendered as one continuous take. |
| resolution | No | 480p, 720p, 1080p | Defaults to 720p. |
| aspectRatio | No | auto, 16:9, 9:16, 1:1, 4:3, 3:4 | auto lets the model frame the shot from the prompt or image. |
| generateAudio | No | boolean | Defaults to true. Dialogue, ambience and effects at no extra cost. |
| imageUrl | No | public URL | Activates image-to-video; the image is the first frame. |
| firstFrameUrl + lastFrameUrl | No | public URLs | Activates first-to-last-frame generation. |
| referenceImageUrls | No | up to 10 public URLs | Activates reference-to-video. Address them by position, e.g. Image 1. |
Polling response
{
"status": "completed",
"url": "https://cdn.imageat.com/generated/video.mp4"
}Production checklist
- Store IMAGEAT_API_KEY only on your server.
- Persist taskId and generationId before polling.
- Use exponential backoff for pending jobs.
- Treat the returned CDN URL as the final MP4 output.
Ready to build with Wan 3.0?
Your existing imageat credits and API keys work here.