img_uuid, and poll or use a callback to receive the generated video.Vidu Q3 Turbo Reference API for Consistent Video Subjects
Access the Vidu Q3 Turbo Reference API through the independent GoEnhance Open API. Send 1–7 reference images with a prompt, persist the returned
Get API Key
Build a valid reference-to-video request
Supply 1–7 reference images
POST to https://api.goenhance.ai/api/v1/videos/generations with model: viduq3_turbo_reference, a required prompt, and an image_urls array containing 1–7 public image URLs.
Track every asynchronous task
Choose a 1–16 second duration and 540p, 720p, or 1080p output. Save the returned img_uuid, then poll GET /api/v1/jobs/detail or configure custom_callback_url.
Validate the reference-only contract before submitting a production task.
Vidu Q3 Turbo Reference API capabilities and parameters
Exact model ID
Set <code>model</code> to <code>viduq3_turbo_reference</code>; preserve the underscore exactly.
Required prompt
Describe the intended action, setting, camera behavior, and visual treatment in <code>prompt</code>.
Reference array
Provide 1–7 public image URLs through <code>image_urls</code>.
Reference-only route
Do not send <code>image_url</code> or <code>image_end_url</code>; use <code>viduq3_turbo</code> for those modes.
Duration
Choose an integer from 1 through 16 seconds. The documented default is 5 seconds.
Resolution
Use <code>540p</code>, <code>720p</code>, or <code>1080p</code>; the documented default is <code>720p</code>.
Estimate total cost as the selected per-second rate multiplied by requested duration.
Vidu Q3 Turbo Reference API pricing
540p
1 token per generated second, equal to $0.02 per second at the documented $0.02-per-token rate.
720p
2.5 tokens per generated second, equal to $0.05 per second at the documented $0.02-per-token rate.
1080p
3.25 tokens per generated second, equal to $0.065 per second at the documented $0.02-per-token rate.
Integrate the Vidu Q3 Turbo Reference API in six steps
1
1. Create an API key
Create a GoEnhance key and send it as a bearer token in the Authorization header.
2
2. Prepare references
Host 1–7 approved reference images at public URLs your request can supply in
image_urls. 3
3. Validate the payload
Set the exact model ID, prompt, duration, and one documented resolution; omit first-frame-only fields.
4
4. Submit the task
POST to
/api/v1/videos/generations and persist the returned img_uuid. 5
5. Track status
Poll
GET /api/v1/jobs/detail or provide a public HTTPS callback URL. 6
6. Handle completion
Process terminal success or failure idempotently and verify the returned video before downstream use.
Build repeatable pipelines where approved visual inputs guide a generated scene.
Developer use cases for reference-driven video
Character continuity
Use several approved character views to guide recurring subjects across short generated clips.
Product storytelling
Provide product references and a scene prompt for reviewable campaign concepts.
Campaign variants
Reuse an approved visual reference set while changing prompts for different placements or story beats.
Previsualization
Turn character, prop, and environment references into motion concepts before full production.
Localized creative
Parameterize scene prompts while keeping the same approved visual references across regional variants.
Serialized content
Guide recurring subjects and art direction across independently submitted asynchronous tasks.
Answers to common input, pricing, task-status, and ShengShu API questions.