> ## Documentation Index
> Fetch the complete documentation index at: https://docs.varios-ai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Google Veo Video

> Video generation through the Gemini API (Veo): async `predictLongRunning`, long-running operation polling, MP4 stored as a persistent resource.

## Options

<ParamField body="model" type="string" required>
  Veo model id (use the identifier from the current [Google Gemini API video / Veo documentation](https://ai.google.dev/gemini-api/docs/video)).
</ParamField>

<ParamField body="googleApiKey" type="string" required>
  Google AI / Gemini API key (`x-goog-api-key`). Store as hidden.
</ParamField>

<ParamField body="endpointUrl" type="string">
  Optional base URL; default `https://generativelanguage.googleapis.com/v1beta/`.
</ParamField>

<ParamField body="costs" type="number">
  Optional flat cost per run (overrides token-based cost).
</ParamField>

<ParamField body="inputTokenCost" type="float">
  Cost per input token (per million) when operation metadata includes token counts.
</ParamField>

<ParamField body="outputTokenCost" type="float">
  Cost per output token (per million) when operation metadata includes token counts.
</ParamField>

## Arguments

<ParamField body="prompt" type="string" required>
  Scene description.
</ParamField>

<ParamField body="aspect_ratio" type="string">
  `16:9`, `9:16`, `1:1`; default `16:9`.
</ParamField>

<ParamField body="duration_seconds" type="integer">
  Duration between 4 and 8 seconds, default 8.
</ParamField>

<ParamField body="resolution" type="string">
  `720p` or `1080p`.
</ParamField>

<ParamField body="negative_prompt" type="string">
  Content to avoid.
</ParamField>

<ParamField body="lastImagesCount" type="integer">
  If greater than 0, uses the latest chat image when present (image-to-video).
</ParamField>

## Prompt function

```
\Avency\BotAPI\Service\Adapter\GoogleVeoVideo::execute
```