Grok Imagine API overview
Grok Imagine is an AI video generation model for short-form creative production. On UniAll AI, the public model id is `grok-imagine`. It supports text-to-video and image-to-video generation through an asynchronous API, making it useful for product clips, ad concepts, social video drafts, storyboards, and automated creative workflows.
The model can generate one video per request with durations from 1 to 15 seconds. Supported aspect ratios include 16:9, 4:3, 3:2, 1:1, 2:3, 3:4, and 9:16, so teams can create both landscape and vertical assets.
Pricing
Grok Imagine is billed per generated second. Current UniAll AI user pricing is:
| Resolution | Price | | --- | ---: | | 480p | $0.034 per second | | 720p | $0.0476 per second |
For example, a 6-second 720p video is estimated at $0.2856 before any account-level adjustments. Failed generations are refundable when the request fails without producing a usable result.
API usage
Use the `/v1/videos/generations` endpoint with `model` set to `grok-imagine`. Requests are asynchronous, so your application should create a generation task and then poll or handle the resulting task status according to your workflow.
```json { "model": "grok-imagine", "generation_mode": "image_to_video", "prompt": "A cinematic product reveal, soft studio lighting, smooth camera movement.", "image_url": "https://example.com/reference.png", "duration": 6, "aspect_ratio": "16:9", "video_count": 1, "resolution": "720p" } ```
For text-to-video, use `generation_mode: "text_to_video"` and omit `image_url`. For image-to-video, provide a PNG, JPEG, or WebP reference image URL.
Who should use it
Grok Imagine is a practical fit for developers, agencies, ecommerce teams, and automation builders who need short video outputs from prompts or reference images. It is especially useful when pricing must be predictable because cost scales directly with duration and resolution.
UniAll AI provides access through a unified API layer, so teams can connect Grok Imagine into existing video pipelines, internal tools, campaign systems, or multi-model generation products without building a separate integration for each model provider.
常见问题
How is the Grok Imagine API priced on UniAll AI?
Grok Imagine is billed per generated second. Current user pricing is $0.034 per second for 480p and $0.0476 per second for 720p.
What is the model id for API requests?
Use the public model id `grok-imagine` in the request body.
What generation modes does Grok Imagine support?
It supports text-to-video and image-to-video generation. Image-to-video accepts one reference image in PNG, JPEG, or WebP format.