start_video_generation
Initiate AI video generation from text prompts, reference images, or frame interpolation using Google's Veo 3.1 model. Returns an operation ID for tracking job completion.
Instructions
Start a Veo 3.1 video generation job. This returns an operation ID immediately - use get_video_job to poll for completion. Supports text-to-video, reference images (up to 3), and first/last frame interpolation.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| prompt | Yes | Text description of the video to generate | |
| model | No | Model to use: veo-3.1-generate-preview (quality, $0.75/sec) or veo-3.1-fast-generate-preview (speed, $0.10/sec). Default: fast | |
| durationSeconds | No | Video duration: 4, 6, or 8 seconds (default: 8) | |
| aspectRatio | No | Aspect ratio (default: 16:9). Note: 9:16 may not work with reference images. | |
| resolution | No | Video resolution (default: 1080p) | |
| seed | No | Optional seed for reproducible generation | |
| sampleCount | No | Number of videos to generate (1-4, default: 1) | |
| generateAudio | No | Whether to generate synchronized audio (default: false). Costs 2x more. | |
| referenceImages | No | Up to 3 reference images for visual guidance. Each can be URL, file path, or fileUri. | |
| firstFrame | No | First frame for interpolation (must also provide lastFrame) | |
| lastFrame | No | Last frame for interpolation (must also provide firstFrame) | |
| negativePrompt | No | Optional: Things to avoid in the video | |
| resizeMode | No | How to fit reference images (default: pad) |