grok-video
Dynamic-duration Grok Imagine video model with text-to-video and image-to-video flows.
向いている用途
Teams that want a dynamic quote model with text and image modes.
What you can send
- Required: model. prompt is required for text-to-video and optional for image-to-video.
- Optional: images (0-7), referenceTaskId, referenceIndex, aspectRatio, duration, resolution, mode, nsfwChecker, projectId
- images and referenceTaskId are mutually exclusive. Choose one source.
Field rules
Up to 7 external image URLs. Cannot be combined with referenceTaskId. Spicy mode is not allowed when external images are provided.
Use a previous Grok task id (from grok-imagine/text-to-image). Cannot be combined with images.
Optional integer 0..5. Selects which of the 6 images from the referenced Grok task to use. Only used with referenceTaskId. Default 0.
Supported: 1:1, 16:9, 2:3, 3:2, 9:16. Default 2:3.
Integer 6..30. Sent as number for T2V, string for I2V.
Supported: 480p, 720p.
fun, normal, or spicy. Default normal. Spicy is not allowed with external images.
Optional boolean.
Pricing formula
Frontend notes
- Show referenceTaskId and referenceIndex only when using a referenced Grok task.
- Show external image picker only when not using a referenced task.
- Hide spicy mode option when external images are present.
- Allow any integer duration from 6 to 30.
- Use callback as the main completion flow.
すぐに試したいですか？
Settings で general API key を作成してから、ここに戻ってリクエスト例をコピーしてください。