Gemini Omni 1.2 Flash

This reference documents the ModelFold gateway.

Endpoint

POST /videos/generations

Full URL: https://modelfold.ai/v1/videos/generations

Authentication: every request must include the Authorization: Bearer <API_KEY> header.

Request Parameters

Body (application/json)

Parameter	Type	Required	Description
model	string	Yes	The model name used to generate the task. This field is required. This interface must use the google/gemini-omni-flash-1-2 model. Default："google/gemini-omni-flash-1-2"。 Allowed values：google/gemini-omni-flash-1-2。 Example：google/gemini-omni-flash-1-2。
prompt	string	Yes	Prompt required by the ModelFold gateway; at most 20000 characters. Video prompt. Required and must be non-empty. Describe the scene, action, camera, and sound. minLength: 1. maxLength: 20000. Example：A cat plays piano in a jazz bar, one continuous shot, soft live jazz。
duration	integer	No	Duration of the generated video in seconds. Send an integer. - Text / reference images (no first frame, no video): required, 3–15 - First frame only: required, 3–15 - First and last frame: required, 3–10 - Video edit (video_list set): do not send; output duration follows the input video ModelFold minimum: 3 seconds; -1 is not supported. minimum: 3. maximum: 15. Example：8。
aspect_ratio	string	No	Aspect ratio of the generated video. Allowed values: 16:9, 9:16, auto. Default 16:9. When no reference image is provided, auto is treated as 16:9. Default："16:9"。 Allowed values：16:9 / 9:16 / auto。 Example：16:9。
resolution	string	No	Resolution of the generated video. Allowed values: 360p, 720p, 1080p, 4k. Default 720p. Default："720p"。 Allowed values：360p / 720p / 1080p / 4k。 Example：720p。
seeds	integer	No	Random seed. Range: 10000–99999. Send a single integer. If omitted, the system generates one. minimum: 10000. maximum: 99999. Example：12345。
image_urls	array	No	Up to 6 publicly accessible reference image URLs, 20 MB each. They are style or content references, not first and last frames. Do not combine with first_frame_url or last_frame_url. Can be combined with video_list.
first_frame_url	string	No	First-frame image URL. Sets the starting frame of the generated video. Use a publicly accessible image URL. - First frame only (no last frame): duration is required, integer 3–15 - Together with last_frame_url: duration is required, integer 3–10 - When this field is set, do not send image_urls or video_list - Do not send last_frame_url alone Example：https://example.com/first.png。
last_frame_url	string	No	Last-frame image URL. Sets the ending frame. Must be sent together with first_frame_url; cannot be used alone. duration is then required and must be 3–10. When first and last frames are set, do not send image_urls or video_list. Example：https://example.com/last.png。
video_list	array<string>	No	Video URL array, max 1. Sending this field selects video-edit mode. Do not send duration; output duration follows the input video. Can be combined with image_urls. Mutually exclusive with first_frame_url and last_frame_url. Use a publicly accessible video URL, MP4 recommended, no longer than 15 seconds. maxItems: 1.

Request Example

curl -X POST "https://modelfold.ai/v1/videos/generations" \
  -H "Authorization: Bearer $MODELFOLD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "google/gemini-omni-flash-1-2",
  "prompt": "A cat plays piano in a jazz bar, one continuous shot, soft live jazz",
  "aspect_ratio": "16:9",
  "resolution": "360p",
  "duration": 3
}'

{
  "model": "google/gemini-omni-flash-1-2",
  "prompt": "A cat plays piano in a jazz bar, one continuous shot, soft live jazz",
  "aspect_ratio": "16:9",
  "resolution": "360p",
  "duration": 3
}

Response

{
  "ok": true,
  "created": 1792134000,
  "data": [
    {
      "url": "https://modelfold.ai/uploads/outputs/result.mp4"
    }
  ],
  "usage": {}
}