Skip to main content
POST
Use Google Veo 3.1 to create videos from images. Supports three modes based on the number of input images.

Available Models

veo3 · veo3_fast · veo3_lite

Generation Modes

Animate a single image as the first frame. The video starts from this image and extends the motion.
Create a smooth transition video between a start frame and an end frame.
Generate a video using 1-3 reference images for style transfer or creative inspiration.
Mode Selection:
  • Auto-inference (default): System infers mode from image count (0→t2v, 1→i2v, 2→i2v-fl, 3→r2v)
  • Explicit mode: Add mode field to override auto-inference (e.g., use mode: "r2v" with 1 image for style reference instead of first-frame control)
I2V vs R2V difference:
  • I2V: Image becomes the video’s first frame (strict frame control)
  • R2V: Image serves as style/content reference (flexible creative generation)
string
required
Model name, e.g. veo3, veo3_fast, veo3_lite.
string
Webhook callback URL.
object
required

Query Result

Image Requirements:
  • Formats: JPEG, PNG, WebP
  • Must be publicly accessible HTTPS URLs
  • Recommended size: < 10MB per image
  • Aspect ratio should match the ratio parameter
Speed Mode Limitations:
  • R2V mode (3 images): Does not support Quality mode. If you use veo3 model, it will automatically fall back to Fast mode. We recommend explicitly using veo3_fast for R2V.
  • Lite mode (veo3_lite): Supports I2V (single image, fixed at 720p) and I2V-FL (first+last frame) only. Does not support R2V — requests with 3 images will fail. The resolution parameter is ignored.