google/veo3.1/reference-to-video

Positive prompt words for content generation (required).

List of reference images for video generation (required). The model will generate a video with a consistent subject based on these images. Accepts 1 to 3 images, which can be provided via URL or Base64 encoding. Supported formats: PNG, JPEG, JPG, WebP. Image dimensions must be at least 128×128 pixels, and each image must not exceed 50MB in size. After Base64 decoding, the length must also not exceed 50MB, and must include the correct content type declaration (such as data:image/jpeg;base64,…).

Resolution for generated video. Optional values: 720p, 1080p.

Whether to generate audio for the video.

Negative prompt words used to exclude or weaken unwanted elements (optional).

The random seed used for generation. A value of -1 indicates random generation. Range: -1 ~ 2147483647.

Callback interface URL (optional). After the task is generated, the system will send a result request to this link. A status code of 200 indicates successful receipt.