Google Veo · Video Generation

Google Veo · IntroductionTemporarily Offline

This model is temporarily offline. The docs are for reference only and calls are unavailable. Watch announcements for restoration.

Google Veo (by Google DeepMind) is a generative AI video model for high-quality, cinematic footage. On this platform we offer the Veo 3.1 family with: • Multimodal input: text-to-video (standard / fast), image-to-video, and first/last-frame interpolation (precise control over opening/closing frames, character & scene transitions). • Output: base duration 4 / 6 / 8s; 720p resolution, optional 1080p (hd=true, extra cost); 16:9 landscape and 9:16 portrait. Returns a taskId; poll for the result.

How to call

  1. Submit: POST /api/cqt/generator/veo, returns a taskId.
  2. Query: GET /api/cqt/info/veo?id=taskId, poll until status = succeeded.

Capability map

Common Errors

CodemsgMeaningHow to fix
451prompt_unsafePrompt rejected by content safety reviewEdit the prompt to remove disallowed content and retry
451video_unsafeThe generated video was flagged as unsafeModify the prompt or change the seed and retry
451image_unsafeInput image flagged as unsafeReplace the input image and retry
400invalid_request / unsupported_*Invalid parameter or unsupported valueCheck model / aspectRatio / duration against the parameter table
500timeoutRequest timed outRetry later; contact support if it persists

Errors are returned as upstream bizCode, e.g. upstream bizCode=451, msg=prompt_unsafe.

Tip: after saving your API Key via Authorize, every capability page supports "Try it".