Google Vertex AI icon

Google Vertex AI

Use Google Vertex AI (Gemini) for text, media analysis, and image generation/editing

Actions11

Video → Generate Video

AI-generated

Summary

Starts an asynchronous Veo video generation job on Google Vertex AI from a text prompt and returns the long-running operation name. It does not return the video itself.

Inputs

  • Video Model (required) — Veo model to use: Veo 2, Veo 3, Veo 3 Fast, Veo 3.1 (default) or Veo 3.1 Fast.
  • Prompt (required) — Text description of the video to generate.
  • Options > Aspect Ratio — 16:9 (default) or 9:16.
  • Options > Duration (Seconds) — Clip length of 4, 6 or 8 seconds. Allowed values vary by model.
  • Options > Generate Audio — Whether to generate audio (Veo 3 and later). Defaults to true.
  • Options > Negative Prompt — What to avoid in the video. Sent only if non-empty.
  • Options > Resolution — 720p (default), 1080p or 4k. 1080p and 4k require an 8s duration and Veo 3.x.
  • Options > Sample Count — Number of videos to generate, 1 to 4.

Output shape

One item per input item, with JSON containing operationName (the Vertex AI long-running operation name) and done (boolean, normally false at start).

No video data is returned. Pass operationName to "Get Video Status" and poll until done is true, then to "Download Video" to get the video as binary. Options are sent only when set, and the API validates model, duration and resolution combinations. Errors are raised as API errors, or returned as an error field if Continue On Fail is on.

Examples

Example 1: Create an 8-second vertical clip with audio from a text prompt

Video Model: Veo 3.1; Prompt: 'A drone shot over a misty forest at sunrise'; Options: Aspect Ratio 9:16, Duration 8, Generate Audio on.

Example 2: Generate several silent draft variations cheaply

Video Model: Veo 3.1 Fast; Options: Sample Count 4, Generate Audio off, Duration 4.

Links

Discussion