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

# Gemini Omni Flash Video Generation

> Generate or edit videos with Gemini Omni Flash using text, up to 10 reference images, or up to 3 input videos

* Use `gemini-omni-flash-preview-official` in the `model` field
* Supports text-to-video, image/reference-to-video, and video editing
* Supports up to 10 reference images or up to 3 input videos; images and videos cannot be mixed
* Video duration is limited to 10 seconds; the Playground offers 4 / 6 / 8 / 10 second presets
* Only `720p` resolution is supported; both `16:9` and `9:16` aspect ratios are available
* Async task API: submit a task, then query by task ID

<Note>
  This page documents `gemini-omni-flash-preview-official`. The platform routes it through the Vertex AI Interactions API with Service Account OAuth, not the Vertex AI Publisher Model `predictLongRunning` endpoint. It remains separate from the legacy `gemini-omni-flash` / `gemini_omni_flash` model and does not change that model's routing or parameters.
</Note>

<Warning>
  Upload local images with the [Upload Image API](../../uploads/images) and local videos with the [Upload Video API](../../uploads/videos) first, then pass the returned URLs. Do not pass base64 media data directly.
</Warning>

## Authorizations

<ParamField header="Authorization" type="string" required>
  Use Bearer Token authentication:

  ```
  Authorization: Bearer YOUR_API_KEY
  ```
</ParamField>

## Body

<ParamField body="model" type="string" default="gemini-omni-flash-preview-official" required>
  Model name. Use `gemini-omni-flash-preview-official`.
</ParamField>

<ParamField body="prompt" type="string" required>
  Text prompt for video generation.
</ParamField>

<ParamField body="duration" type="integer" default="6">
  Video duration in seconds. The maximum is `10`; the Playground offers `4`, `6`, `8`, and `10` second presets.
</ParamField>

<ParamField body="aspect_ratio" type="string" default="16:9">
  Video aspect ratio:

  * `16:9` landscape
  * `9:16` portrait
</ParamField>

<ParamField body="resolution" type="string" default="720p">
  Video resolution:

  * `720p` default resolution
  * Only `720p` is supported; Veo-specific parameters are not accepted
</ParamField>

<ParamField body="image_urls" type="string[]">
  Optional reference image URL array with up to `10` items. Omit it for text-to-video. This field cannot be combined with `video_list`.
</ParamField>

<ParamField body="video_list" type="object[]">
  Optional input videos for video editing, with up to `3` items. Each object must contain `video_url`, for example `{ "video_url": "https://example.com/input.mp4" }`. HTTPS URLs, data URIs, and platform-uploaded video URLs are supported. Each downloaded input video is limited to 50 MB. This field cannot be combined with `image_urls`.
</ParamField>

<ParamField body="metadata" type="object">
  Optional Interactions generation controls. Supported keys include `temperature` (`0.0`–`2.0`), `topP` / `top_p` (`0.0`–`1.0`), `candidateCount` / `candidate_count`, `previous_interaction_id`, and `task` (`text_to_video`, `image_to_video`, `reference_to_video`, or `edit`). When `task` is omitted, the platform does not invent a default; the upstream model infers the mode from the prompt and media. For `edit` (or video inputs without an explicit non-edit task), do not rely on `aspect_ratio`; the platform omits it for that path because upstream rejects aspect ratio on edit tasks.
</ParamField>

## Response

<ResponseField name="id" type="string">
  Task ID for status polling.
</ResponseField>

<ResponseField name="object" type="string">
  Object type, usually `generation.task`.
</ResponseField>

<ResponseField name="model" type="string">
  Model name used for the request.
</ResponseField>

<ResponseField name="status" type="string">
  Task status: `queued`, `in_progress`, `completed`, or `failed`.
</ResponseField>

<ResponseField name="created_at" type="integer">
  Task creation timestamp.
</ResponseField>

<ResponseField name="video_url" type="string">
  Generated video URL when the task completes.
</ResponseField>

## Examples

<RequestExample>
  ```bash cURL theme={null}
  curl --request POST \
    --url https://toapis.com/v1/videos/generations \
    --header 'Authorization: Bearer YOUR_API_KEY' \
    --header 'Content-Type: application/json' \
    --data '{
      "model": "gemini-omni-flash-preview-official",
      "prompt": "A blue pump bottle slowly rotates in a clean white studio, commercial product video",
      "duration": 6,
      "aspect_ratio": "9:16",
      "resolution": "720p",
      "image_urls": ["https://example.com/reference.jpg"]
    }'
  ```
</RequestExample>

Video editing example:

```bash theme={null}
curl --request POST \
  --url https://toapis.com/v1/videos/generations \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "gemini-omni-flash-preview-official",
    "prompt": "Turn the daytime scene into a rainy night while preserving the camera motion",
    "duration": 8,
    "aspect_ratio": "16:9",
    "resolution": "720p",
    "video_list": [{"video_url": "https://example.com/input.mp4"}],
    "metadata": {"temperature": 1.0, "topP": 0.95}
  }'
```

<ResponseExample>
  ```json 200 theme={null}
  {
    "id": "video_01JZEXAMPLE",
    "object": "generation.task",
    "model": "gemini-omni-flash-preview-official",
    "status": "queued",
    "created_at": 1779247407
  }
  ```
</ResponseExample>

## Query Task

The generation endpoint returns a task ID. Use the common video task status endpoint to fetch status and results:

```bash theme={null}
curl --request GET \
  --url https://toapis.com/v1/videos/generations/{task_id} \
  --header "Authorization: Bearer YOUR_API_KEY"
```