Pixazo API→Models→Text to Video API→Gemini Omni 1.2

Gemini Omni 1.2 Flash 1.2 Flash API: Pricing, Documentation

Google's Gemini Omni 1.2 Flash is a multimodal video model that handles text-to-video, image-to-video, reference-to-video, and video-to-video editing through a single API. Clip length is selectable from 3 to 15 seconds via the duration parameter; video-to-video edits inherit the source clip's length. Successor to Gemini Omni 1.1 Flash, with stronger character consistency and lip-sync.

Get API Key

Models Version

WELCOME BONUS

Get 10% Cashback on Your First Payment (up to $500)

Add funds and get 10% cashback on your first payment (up to $500).

Claim Your Cashback →

Gemini Omni 1.2 Flash API Documentation

Generate a video from a text prompt only. Asynchronous: submit returns a request_id; poll the status endpoint until the request is COMPLETED. Typical generation time is ~36s.

Backed by Google's gemini-omni-1.2-flash.

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video

Authentication

All requests require an API key passed via header.

Header	Type	Required	Description
Ocp-Apim-Subscription-Key	string	Yes	Your API subscription key

Text to Video - Gemini Omni 1.2 Flash

Request Code

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video
Content-Type: application/json
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.",
  "aspect_ratio": "16:9"
}

import requests

url = "https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video"
headers = {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
  "prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.",
  "aspect_ratio": "16:9"
}

resp = requests.post(url, json=data, headers=headers)
print(resp.json())

const res = await fetch("https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
  },
  body: JSON.stringify({
  "prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.",
  "aspect_ratio": "16:9"
})
});
console.log(await res.json());

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  --data-raw '{"prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.", "aspect_ratio": "16:9"}'

Output

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Try Now

Webhook (Optional)

Instead of polling, you can receive a webhook callback when the request reaches a terminal state. Provide a webhook URL via header on the submit request.

Header	Required	Description
`X-Webhook-URL`	Yes (to enable)	HTTPS endpoint on your server that will receive the `POST` callback. Must respond `2xx` within a few seconds.
`X-Webhook-Mode`	No	`terminal` (default) — fires once at the final status (`COMPLETED`/`FAILED`/`ERROR`). `sync` — fires on every poll cycle plus the terminal event, and caps the queue's polling delay at 15s for tighter progress updates.

Example: enable webhook

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/text-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook/callback' \
  --data-raw '{"prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.", "aspect_ratio": "16:9"}'

Callback Payload

Your endpoint receives a POST application/json with the same shape as the GET /v2/requests/status/{request_id} response. Example terminal callback (mode terminal):

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Failure callback shape

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "gemini-omni-1-2",
  "error": "Description of the failure",
  "output": null,
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:22",
  "completed_at": null
}

Delivery semantics

terminal mode: one webhook callback when the request is COMPLETED or ERROR/FAILED.
sync mode: a webhook callback on each status change.
Callbacks are idempotent on request_id — de-duplicate on it.
Respond 200 within a few seconds; the webhook endpoint must be HTTPS.

Request Parameters

Parameter	Required	Type	Default	Allowed values / range	Description
`prompt`	Yes	string	—	—	Text prompt describing the video to generate.
`aspect_ratio`	No	string	`16:9`	`16:9`, `9:16`	Output video aspect ratio.
`duration`	No	integer	—	3–15 (integer seconds)	Length of the generated video, in seconds. Accepted values are integers 3–15. Optional — if you omit it, the video is 10 seconds long. When you send it, it takes precedence over any length mentioned in `prompt`. Billing follows the actual output length, so a shorter clip costs proportionally less.

Example Request

{
  "prompt": "A golden retriever running through a sunlit meadow, slow motion, cinematic.",
  "aspect_ratio": "16:9"
}

Example Response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Request Headers

Header	Required	Description
Content-Type	Yes	`application/json`
Ocp-Apim-Subscription-Key	Yes	Your API subscription key.
X-Webhook-URL	No	Enable webhook callbacks (see Webhook section).

Response Handling

Status Code	Meaning
202	Accepted — request queued; returns `request_id` and `polling_url`.
400	Bad request — invalid or missing parameters (e.g. missing `prompt`, missing source image/video).
401	Unauthorized — missing or invalid subscription key.
402	Insufficient balance.
403	Forbidden.
429	Too many requests.
500	Internal server error.

Error Responses

{
  "error": "Insufficient Balance",
  "message": "Your wallet does not have enough balance."
}

A failed generation is reported via the status endpoint (or webhook) with status: "FAILED" or "ERROR" and an error message. Malformed requests (e.g. a missing required field) are rejected immediately with 400.

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR). Typical generation time is ~36s.

curl 'https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Response Fields

Field	Type	Description
request_id	string	Unique request identifier.
status	string	`QUEUED`, `PROCESSING`, `COMPLETED`, `FAILED`, or `ERROR`.
model_id	string	The model that handled the request (`gemini-omni-1-2`).
output.media_url	array	URL(s) to the generated MP4 video.
output.media_type	string	Always `video/mp4`.
created_at	string	Request creation timestamp.
completed_at	string	Completion timestamp.
polling_url	string	Status URL for this request (submit response only).
error	string	Error message when `status` is `FAILED`/`ERROR`.

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed per second of output video, measured from the generated file, so a shorter clip costs proportionally less. Failed requests are not billed. The current rate is shown on this model's page.

Notes & Limitations

Topic	Detail
Audio	Generated automatically for every video — there is no on/off parameter. Describe the sound effects, music, or dialogue you want directly in `prompt`.
Prompt controls	No separate negative-prompt, temperature, or sampling parameter. Put exclusions directly in `prompt` (e.g. “no dialogue”). See the Prompting Guide below.
Language	English is fully supported; other languages may work but results can vary.
Generation time	Varies with duration, resolution, and current API load — longer or higher-resolution videos take longer.
Watermark	Every generated video carries an invisible SynthID watermark (not visible to viewers, but programmatically detectable).
Content safety	Filters apply to both your prompt and the generated video, and enforcement is region-dependent. Policy-violating prompts or content are blocked.
Not supported	Uploading audio references is not supported in the current API version.

Prompting Guide

The prompt is passed through to the model verbatim, so all of Google’s prompt syntax works as-is.

Technique	How to use it
Single scene	By default the model may split into several shots. Force one continuous shot by saying so — “in a single continuous shot”, “no scene cuts”.
Remove elements	Use simple negatives in the prompt — “no dialogue”, “no embellishments”, “no extra sound effects”.
Prompt the audio	Describe the sound you want — “calm background music”, “a high-energy techno beat”, “a low tinny radio in the background”.
Timing events	Use natural language (“after 3 seconds a woman enters”) or timecode syntax: `[0-5s] walking [5-10s] stops and turns [10-15s] starts running`.
On-screen text	Quote it exactly — e.g. a street sign that says: “This is an AI generation”.
Meta-prompting	Ask for general qualities — “be extremely detailed in your descriptions of characters and environments”.

Gemini Omni 1.2 Flash API Pricing

Resolution	Price (USD)
Per second of output video	$0.07297

Image to Video

Gemini Omni 1.2 Flash API Documentation

Animate a single source image into a video, guided by a text prompt. Asynchronous: submit returns a request_id; poll the status endpoint until the request is COMPLETED. Typical generation time is ~36s.

Backed by Google's gemini-omni-1.2-flash.

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video

Authentication

All requests require an API key passed via header.

Header	Type	Required	Description
Ocp-Apim-Subscription-Key	string	Yes	Your API subscription key

Image to Video - Gemini Omni 1.2 Flash

Request Code

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video
Content-Type: application/json
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.",
  "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/nano-banana_019f1840-7f48-76a2-3393-d5f174826bae8/output.jpg",
  "aspect_ratio": "9:16"
}

import requests

url = "https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video"
headers = {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
  "prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.",
  "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/nano-banana_019f1840-7f48-76a2-3393-d5f174826bae8/output.jpg",
  "aspect_ratio": "9:16"
}

resp = requests.post(url, json=data, headers=headers)
print(resp.json())

const res = await fetch("https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
  },
  body: JSON.stringify({
  "prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.",
  "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/nano-banana_019f1840-7f48-76a2-3393-d5f174826bae8/output.jpg",
  "aspect_ratio": "9:16"
})
});
console.log(await res.json());

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  --data-raw '{"prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.", "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/nano-banana_019f1840-7f48-76a2-3393-d5f174826bae8/output.jpg", "aspect_ratio": "9:16"}'

Output

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Try Now

Webhook (Optional)

Instead of polling, you can receive a webhook callback when the request reaches a terminal state. Provide a webhook URL via header on the submit request.

Header	Required	Description
`X-Webhook-URL`	Yes (to enable)	HTTPS endpoint on your server that will receive the `POST` callback. Must respond `2xx` within a few seconds.
`X-Webhook-Mode`	No	`terminal` (default) — fires once at the final status (`COMPLETED`/`FAILED`/`ERROR`). `sync` — fires on every poll cycle plus the terminal event, and caps the queue's polling delay at 15s for tighter progress updates.

Example: enable webhook

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/image-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook/callback' \
  --data-raw '{"prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.", "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg", "aspect_ratio": "9:16"}'

Callback Payload

Your endpoint receives a POST application/json with the same shape as the GET /v2/requests/status/{request_id} response. Example terminal callback (mode terminal):

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Failure callback shape

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "gemini-omni-1-2",
  "error": "Description of the failure",
  "output": null,
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:22",
  "completed_at": null
}

Delivery semantics

terminal mode: one webhook callback when the request is COMPLETED or ERROR/FAILED.
sync mode: a webhook callback on each status change.
Callbacks are idempotent on request_id — de-duplicate on it.
Respond 200 within a few seconds; the webhook endpoint must be HTTPS.

Request Parameters

Parameter	Required	Type	Default	Allowed values / range	Description
`prompt`	Yes	string	—	—	Text prompt describing the motion/scene to animate the image into.
`image_url`	Yes	string (URL)	—	JPEG, PNG, WebP, HEIC, or HEIF	URL of the source image to animate.
`image_urls`	No	string[] (URLs)	—	JPEG, PNG, WebP, HEIC, or HEIF	Alternative to `image_url`. Only the first entry is used by this operation — to combine several images in one video, use the Reference to Video operation instead.
`aspect_ratio`	No	string	`16:9`	`16:9`, `9:16`	Output video aspect ratio.
`duration`	No	integer	—	3–15 (integer seconds)	Length of the generated video, in seconds. Accepted values are integers 3–15. Optional — if you omit it, the video is 10 seconds long. When you send it, it takes precedence over any length mentioned in `prompt`. Billing follows the actual output length, so a shorter clip costs proportionally less.

Content Item Types & Limits

Type	Max	Format / Size	Description
image	1	JPG, PNG, WEBP, HEIC, HEIF · < 20 MB	Input image to animate.

Example Request

{
  "prompt": "Animate the photo with a gentle camera pan and the subject blinking naturally.",
  "image_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/nano-banana_019f1840-7f48-76a2-3393-d5f174826bae8/output.jpg",
  "aspect_ratio": "9:16"
}

Example Response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Request Headers

Header	Required	Description
Content-Type	Yes	`application/json`
Ocp-Apim-Subscription-Key	Yes	Your API subscription key.
X-Webhook-URL	No	Enable webhook callbacks (see Webhook section).

Response Handling

Status Code	Meaning
202	Accepted — request queued; returns `request_id` and `polling_url`.
400	Bad request — invalid or missing parameters (e.g. missing `prompt`, missing source image/video).
401	Unauthorized — missing or invalid subscription key.
402	Insufficient balance.
403	Forbidden.
429	Too many requests.
500	Internal server error.

Error Responses

{
  "error": "Insufficient Balance",
  "message": "Your wallet does not have enough balance."
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR). Typical generation time is ~36s.

curl 'https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Response Fields

Field	Type	Description
request_id	string	Unique request identifier.
status	string	`QUEUED`, `PROCESSING`, `COMPLETED`, `FAILED`, or `ERROR`.
model_id	string	The model that handled the request (`gemini-omni-1-2`).
output.media_url	array	URL(s) to the generated MP4 video.
output.media_type	string	Always `video/mp4`.
created_at	string	Request creation timestamp.
completed_at	string	Completion timestamp.
polling_url	string	Status URL for this request (submit response only).
error	string	Error message when `status` is `FAILED`/`ERROR`.

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed per second of output video, measured from the generated file, so a shorter clip costs proportionally less. Failed requests are not billed. The current rate is shown on this model's page.

Notes & Limitations

Topic	Detail
Input image	JPEG, PNG, WebP, HEIC, or HEIF. Keep the combined inline image data under ~20 MB (Google’s inline-data limit).
Audio	Generated automatically for every video — there is no on/off parameter. Describe the sound effects, music, or dialogue you want directly in `prompt`.
Prompt controls	No separate negative-prompt, temperature, or sampling parameter. Put exclusions directly in `prompt` (e.g. “no dialogue”). See the Prompting Guide below.
Language	English is fully supported; other languages may work but results can vary.
Generation time	Varies with duration, resolution, and current API load — longer or higher-resolution videos take longer.
Watermark	Every generated video carries an invisible SynthID watermark (not visible to viewers, but programmatically detectable).
Content safety	Filters apply to both your prompt and the generated video, and enforcement is region-dependent. Policy-violating prompts or content are blocked.
Not supported	Uploading audio references is not supported in the current API version.
Regional limits	In the EEA, Switzerland, and the UK, images containing minors cannot be used. Images of certain recognizable (real, identifiable) people are not supported anywhere.

Prompting Guide

The prompt is passed through to the model verbatim, so all of Google’s prompt syntax works as-is.

Technique	How to use it
First frame	The image is used as the start frame. Say so in plain language when the prompt is complex — e.g. `starting from the image, a woman starts walking`.
No control tags	Do not put control strings such as `<FIRST_FRAME>` or `<IMAGE_REF_0>` in `prompt`: Gemini Omni 1.2 Flash no longer supports them. Describe in plain language how each image should be used instead.
Single scene	By default the model may split into several shots. Force one continuous shot by saying so — “in a single continuous shot”, “no scene cuts”.
Remove elements	Use simple negatives in the prompt — “no dialogue”, “no embellishments”, “no extra sound effects”.
Prompt the audio	Describe the sound you want — “calm background music”, “a high-energy techno beat”, “a low tinny radio in the background”.
Timing events	Use natural language (“after 3 seconds a woman enters”) or timecode syntax: `[0-5s] walking [5-10s] stops and turns [10-15s] starts running`.
On-screen text	Quote it exactly — e.g. a street sign that says: “This is an AI generation”.
Meta-prompting	Ask for general qualities — “be extremely detailed in your descriptions of characters and environments”.

Gemini Omni 1.2 Flash API Pricing

Resolution	Price (USD)
Per second of output video	$0.07297

Reference to Video (Ref Images + Ref Videos to Video)

Gemini Omni 1.2 Flash API Documentation

Generate a video guided by reference images, short reference videos, or both, plus a text prompt. Asynchronous: submit returns a request_id; poll the status endpoint until the request is COMPLETED. Typical generation time is ~36s with reference images only; reference videos add upload time (our 3-second test with one clip took about 1 min 45 s).

Backed by Google's gemini-omni-1.2-flash.

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video

Authentication

All requests require an API key passed via header.

Header	Type	Required	Description
Ocp-Apim-Subscription-Key	string	Yes	Your API subscription key

Reference to Video (Ref Images + Ref Videos to Video) - Gemini Omni 1.2 Flash

Request Code

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video
Content-Type: application/json
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "prompt": "The woman from the reference image meets the subject shown in the reference video on a bright city street, in a single continuous shot.",
  "image_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg"
  ],
  "video_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-3s.mp4"
  ],
  "duration": 5,
  "aspect_ratio": "16:9"
}

import requests

url = "https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video"
headers = {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
  "prompt": "The woman from the reference image meets the subject shown in the reference video on a bright city street, in a single continuous shot.",
  "image_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg"
  ],
  "video_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-3s.mp4"
  ],
  "duration": 5,
  "aspect_ratio": "16:9"
}

resp = requests.post(url, json=data, headers=headers)
print(resp.json())

const res = await fetch("https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
  },
  body: JSON.stringify({
  "prompt": "The woman from the reference image meets the subject shown in the reference video on a bright city street, in a single continuous shot.",
  "image_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg"
  ],
  "video_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-3s.mp4"
  ],
  "duration": 5,
  "aspect_ratio": "16:9"
})
});
console.log(await res.json());

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  --data-raw '{"prompt": "The woman from the reference image meets the subject shown in the reference video on a bright city street, in a single continuous shot.", "image_urls": ["https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg"], "video_urls": ["https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-3s.mp4"], "duration": 5, "aspect_ratio": "16:9"}'

Output

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Try Now

Webhook (Optional)

Instead of polling, you can receive a webhook callback when the request reaches a terminal state. Provide a webhook URL via header on the submit request.

Header	Required	Description
`X-Webhook-URL`	Yes (to enable)	HTTPS endpoint on your server that will receive the `POST` callback. Must respond `2xx` within a few seconds.
`X-Webhook-Mode`	No	`terminal` (default) — fires once at the final status (`COMPLETED`/`FAILED`/`ERROR`). `sync` — fires on every poll cycle plus the terminal event, and caps the queue's polling delay at 15s for tighter progress updates.

Example: enable webhook

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/reference-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook/callback' \
  --data-raw '{"prompt": "The woman from the reference image meets the subject shown in the reference video on a bright city street, in a single continuous shot.", "image_urls": ["https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/images/input.jpg"], "video_urls": ["https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-3s.mp4"], "duration": 5, "aspect_ratio": "16:9"}'

Callback Payload

Your endpoint receives a POST application/json with the same shape as the GET /v2/requests/status/{request_id} response. Example terminal callback (mode terminal):

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Failure callback shape

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "gemini-omni-1-2",
  "error": "Description of the failure",
  "output": null,
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:22",
  "completed_at": null
}

Delivery semantics

terminal mode: one webhook callback when the request is COMPLETED or ERROR/FAILED.
sync mode: a webhook callback on each status change.
Callbacks are idempotent on request_id — de-duplicate on it.
Respond 200 within a few seconds; the webhook endpoint must be HTTPS.

Request Parameters

Parameter	Required	Type	Default	Allowed values / range	Description
`prompt`	Yes	string	—	—	Text prompt describing the scene to generate from the reference images and videos.
`image_urls`	Yes, unless `video_urls` is sent	string[] (URLs)	—	JPEG, PNG, WebP, HEIC, or HEIF	Reference images to incorporate into the video, up to 6 (up to 3 images per character, for up to 2 characters). Say in plain language in `prompt` how each image is used, e.g. “the woman from the first image”.
`image_url`	No	string (URL)	—	JPEG, PNG, WebP, HEIC, or HEIF	Convenience alternative to `image_urls` when you have a single reference image; it is treated as one reference.
`video_urls`	No	string[] (https URLs)	—	MP4 or MOV · up to 3 seconds and 50 MB each · 1–3 clips	Short reference videos of a person, character, or object the video should feature; Google says they work best for keeping a visual likeness. Up to 3 when `duration` is 3–10 (or left out), 1 when it is 11–15. Each must be a public https URL to a regular (not fragmented) MP4 or MOV file. Their sound is ignored. Say in plain language in `prompt` how each clip is used, e.g. “the man shown in the reference video”. A clip that is longer than 3 seconds, larger than 50 MB, cannot be read, or is one too many is refused with a `400` before anything is charged.
`aspect_ratio`	No	string	`16:9`	`16:9`, `9:16`	Output video aspect ratio.
`duration`	No	integer	—	3–15 (integer seconds)	Length of the generated video, in seconds. Accepted values are integers 3–15. Optional — if you omit it, the video is 10 seconds long. When you send it, it takes precedence over any length mentioned in `prompt`. Billing follows the actual output length, so a shorter clip costs proportionally less.

Content Item Types & Limits

Type	Max	Format / Size	Description
image	6	JPG, PNG, WEBP, HEIC, HEIF · < 20 MB combined	Reference images (`image_urls` / `image_url`): up to 6, up to 3 per character for up to 2 characters.
video	3 (1 when `duration` is 11–15)	MP4, MOV · up to 3 s · up to 50 MB each	Reference videos (`video_urls`). Audio in them is ignored.

Example Request

{
  "prompt": "Transform both robots into a sleek futuristic sports car through a seamless mechanical transformation, with shifting metal panels, rotating gears, glowing energy, and cinematic motion.",
  "image_urls": [
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/nano-banana/nano-banana-a382a80b-f8df-4de1-a0c1-a5dcfd42dae4-1758783383399.jpg",
    "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/flux-schnell-cf/prompt-1782482807515-639182.png"
  ],
  "aspect_ratio": "16:9"
}

Example Response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Request Headers

Header	Required	Description
Content-Type	Yes	`application/json`
Ocp-Apim-Subscription-Key	Yes	Your API subscription key.
X-Webhook-URL	No	Enable webhook callbacks (see Webhook section).

Response Handling

Status Code	Meaning
202	Accepted — request queued; returns `request_id` and `polling_url`.
400	Bad request — invalid or missing parameters (e.g. missing `prompt`, missing source image/video, a reference video longer than 3 seconds or larger than 50 MB, or more reference videos than the duration allows).
401	Unauthorized — missing or invalid subscription key.
402	Insufficient balance.
403	Forbidden.
429	Too many requests.
500	Internal server error.

Error Responses

{
  "error": "Insufficient Balance",
  "message": "Your wallet does not have enough balance."
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR). Typical generation time is ~36s with reference images only; reference videos add upload time (our 3-second test with one clip took about 1 min 45 s).

curl 'https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Response Fields

Field	Type	Description
request_id	string	Unique request identifier.
status	string	`QUEUED`, `PROCESSING`, `COMPLETED`, `FAILED`, or `ERROR`.
model_id	string	The model that handled the request (`gemini-omni-1-2`).
output.media_url	array	URL(s) to the generated MP4 video.
output.media_type	string	Always `video/mp4`.
created_at	string	Request creation timestamp.
completed_at	string	Completion timestamp.
polling_url	string	Status URL for this request (submit response only).
error	string	Error message when `status` is `FAILED`/`ERROR`.

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed per second of output video, measured from the generated file, so a shorter clip costs proportionally less. Failed requests are not billed. The current rate is shown on this model's page.

Notes & Limitations

Topic	Detail
Input images	JPEG, PNG, WebP, HEIC, or HEIF. Keep the combined inline image data under ~20 MB (Google’s inline-data limit).
References	Up to 6 reference images per request (up to 3 images per character, for up to 2 characters), and short reference videos in `video_urls`: MP4 or MOV, up to 3 seconds and 50 MB each, up to 3 per request for a 3–10 second video and 1 for an 11–15 second video. A request may carry images, videos, or both. The sound of a reference video is ignored. Google notes that reasoning across several videos is not supported and may give unexpected results, so one clip is the most reliable. The price is per second of output video; reference images and videos do not change it.
Audio	Generated automatically for every video — there is no on/off parameter. Describe the sound effects, music, or dialogue you want directly in `prompt`.
Prompt controls	No separate negative-prompt, temperature, or sampling parameter. Put exclusions directly in `prompt` (e.g. “no dialogue”). See the Prompting Guide below.
Language	English is fully supported; other languages may work but results can vary.
Generation time	Varies with duration, resolution, and current API load — longer or higher-resolution videos take longer. Reference videos add upload and processing time before generation starts.
Watermark	Every generated video carries an invisible SynthID watermark (not visible to viewers, but programmatically detectable).
Content safety	Filters apply to both your prompt and the generated video, and enforcement is region-dependent. Policy-violating prompts or content are blocked.
Not supported	Uploading audio references is not supported in the current API version. Reference videos are accepted only here (Reference to Video), not by Video to Video editing.
Regional limits	In the EEA, Switzerland, and the UK, images containing minors cannot be used. Images of certain recognizable (real, identifiable) people are not supported anywhere.

Prompting Guide

The prompt is passed through to the model verbatim, so all of Google’s prompt syntax works as-is.

Technique	How to use it
Referring to images and videos	Name each image by its place in `image_urls` and its role, in plain language — e.g. `in the style of the first image, the woman from the second image is walking`. Name a reference video the same way — e.g. `the man shown in the reference video sits at the cafe table from the first image`.
No control tags	Do not put control strings such as `<FIRST_FRAME>`, `<IMAGE_REF_0>` or `<VIDEO_REF_0>` in `prompt`: Gemini Omni 1.2 Flash no longer supports them. Describe in plain language how each image or video should be used instead.
Multi-image compositions	Combine plain-language references with timecodes — e.g. `[0-5s] the woman from the first image holds the bag from the second image [5-10s] the man from the third image holds the hat from the fourth image`.
Single scene	By default the model may split into several shots. Force one continuous shot by saying so — “in a single continuous shot”, “no scene cuts”.
Remove elements	Use simple negatives in the prompt — “no dialogue”, “no embellishments”, “no extra sound effects”.
Prompt the audio	Describe the sound you want — “calm background music”, “a high-energy techno beat”, “a low tinny radio in the background”.
Timing events	Use natural language (“after 3 seconds a woman enters”) or timecode syntax: `[0-5s] walking [5-10s] stops and turns [10-15s] starts running`.
On-screen text	Quote it exactly — e.g. a street sign that says: “This is an AI generation”.
Meta-prompting	Ask for general qualities — “be extremely detailed in your descriptions of characters and environments”.

Gemini Omni 1.2 Flash API Pricing

Resolution	Price (USD)
Per second of output video	$0.07297

Video to Video(video editing)

Gemini Omni 1.2 Flash API Documentation

Edit/transform an existing video using a text prompt (e.g. background swap, style change, object edits). The output inherits the source video's dimensions. Asynchronous: submit returns a request_id; poll the status endpoint until the request is COMPLETED. Typical generation time is ~60-110s.

Backed by Google's gemini-omni-1.2-flash.

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video

Authentication

All requests require an API key passed via header.

Header	Type	Required	Description
Ocp-Apim-Subscription-Key	string	Yes	Your API subscription key

Video to Video - Gemini Omni 1.2 Flash

Request Code

POST https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video
Content-Type: application/json
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
  "prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.",
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-8s.mp4"
}

import requests

url = "https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video"
headers = {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
}
data = {
  "prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.",
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-8s.mp4"
}

resp = requests.post(url, json=data, headers=headers)
print(resp.json())

const res = await fetch("https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Ocp-Apim-Subscription-Key": "YOUR_SUBSCRIPTION_KEY"
  },
  body: JSON.stringify({
  "prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.",
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-8s.mp4"
})
});
console.log(await res.json());

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  --data-raw '{"prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.", "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-8s.mp4"}'

Output

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Try Now

Webhook (Optional)

Instead of polling, you can receive a webhook callback when the request reaches a terminal state. Provide a webhook URL via header on the submit request.

Header	Required	Description
`X-Webhook-URL`	Yes (to enable)	HTTPS endpoint on your server that will receive the `POST` callback. Must respond `2xx` within a few seconds.
`X-Webhook-Mode`	No	`terminal` (default) — fires once at the final status (`COMPLETED`/`FAILED`/`ERROR`). `sync` — fires on every poll cycle plus the terminal event, and caps the queue's polling delay at 15s for tighter progress updates.

Example: enable webhook

curl -X POST 'https://gateway.pixazo.ai/gemini-omni-1-2/v1/video-to-video' \
  -H 'Content-Type: application/json' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY' \
  -H 'X-Webhook-URL: https://your-server.com/webhook/callback' \
  --data-raw '{"prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.", "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/doc-assets/video/input-8s.mp4"}'

Callback Payload

Your endpoint receives a POST application/json with the same shape as the GET /v2/requests/status/{request_id} response. Example terminal callback (mode terminal):

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Failure callback shape

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "ERROR",
  "model_id": "gemini-omni-1-2",
  "error": "Description of the failure",
  "output": null,
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:22",
  "completed_at": null
}

Delivery semantics

terminal mode: one webhook callback when the request is COMPLETED or ERROR/FAILED.
sync mode: a webhook callback on each status change.
Callbacks are idempotent on request_id — de-duplicate on it.
Respond 200 within a few seconds; the webhook endpoint must be HTTPS.

Request Parameters

Parameter	Required	Type	Default	Allowed values / range	Description
`prompt`	Yes	string	—	—	Text describing the edit to apply.
`video_url`	Yes	string (URL)	—	MP4, MPEG, MOV, AVI, FLV, MPG, WebM, WMV, or 3GPP	Source video to edit. Must be a direct, fetchable video file URL — YouTube links and other hosted-platform URLs are not supported. The source clip must be 15 seconds or shorter.
`aspect_ratio`	No	string	—	—	Not used for video-to-video — the edited clip keeps the aspect ratio of the source video.
`duration`	No	integer	—	—	Not used for video-to-video — the edited clip inherits the length of the source video. Billing follows the actual output length.

Content Item Types & Limits

Type	Max	Format / Size	Description
video	1	MP4, MOV, WEBM, AVI, MPEG, FLV, 3GP	Source video to transform.

Example Request

{
  "prompt": "Replace the background with a neon cyberpunk city while keeping the subject unchanged.",
  "video_url": "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/sync-lipsync-v2-pro_019f1910-7eca-7dc9-c2ef-da59f4839abec/output.mp4"
}

Example Response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "QUEUED",
  "polling_url": "https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
}

Request Headers

Header	Required	Description
Content-Type	Yes	`application/json`
Ocp-Apim-Subscription-Key	Yes	Your API subscription key.
X-Webhook-URL	No	Enable webhook callbacks (see Webhook section).

Response Handling

Status Code	Meaning
202	Accepted — request queued; returns `request_id` and `polling_url`.
400	Bad request — invalid or missing parameters (e.g. missing `prompt`, missing source image/video).
401	Unauthorized — missing or invalid subscription key.
402	Insufficient balance.
403	Forbidden.
429	Too many requests.
500	Internal server error.

Error Responses

{
  "error": "Insufficient Balance",
  "message": "Your wallet does not have enough balance."
}

Retrieving Results

Poll the status endpoint with the request_id from the submit response until status is COMPLETED (or FAILED/ERROR). Typical generation time is ~60-110s.

curl 'https://gateway.pixazo.ai/v2/requests/status/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx' \
  -H 'Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY'

Completed response

{
  "request_id": "gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "status": "COMPLETED",
  "model_id": "gemini-omni-1-2",
  "error": null,
  "output": {
    "media_url": [
      "https://pub-582b7213209642b9b995c96c95a30381.r2.dev/v1/gemini-omni-1-2_01a0xxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/output.mp4"
    ],
    "media_type": "video/mp4"
  },
  "created_at": "2026-07-01T10:00:00.000Z",
  "updated_at": "2026-07-01 10:00:48",
  "completed_at": "2026-07-01 10:00:48"
}

Response Fields

Field	Type	Description
request_id	string	Unique request identifier.
status	string	`QUEUED`, `PROCESSING`, `COMPLETED`, `FAILED`, or `ERROR`.
model_id	string	The model that handled the request (`gemini-omni-1-2`).
output.media_url	array	URL(s) to the generated MP4 video.
output.media_type	string	Always `video/mp4`.
created_at	string	Request creation timestamp.
completed_at	string	Completion timestamp.
polling_url	string	Status URL for this request (submit response only).
error	string	Error message when `status` is `FAILED`/`ERROR`.

Status Values & Flow

QUEUED → PROCESSING → COMPLETED (success) or FAILED/ERROR (failure).

Pricing

Billed per second of output video, measured from the generated file, so a shorter clip costs proportionally less. Failed requests are not billed. The current rate is shown on this model's page.

Notes & Limitations

Topic	Detail
Input video	MP4, MPEG, MOV, AVI, FLV, MPG, WebM, WMV, or 3GPP. Use short, modestly-sized clips for best reliability — very large or long videos may be slow or fail.
Audio	Generated automatically for every video — there is no on/off parameter. Describe the sound effects, music, or dialogue you want directly in `prompt`.
Prompt controls	No separate negative-prompt, temperature, or sampling parameter. Put exclusions directly in `prompt` (e.g. “no dialogue”). See the Prompting Guide below.
Language	English is fully supported; other languages may work but results can vary.
Generation time	Varies with duration, resolution, and current API load — longer or higher-resolution videos take longer.
Watermark	Every generated video carries an invisible SynthID watermark (not visible to viewers, but programmatically detectable).
Content safety	Filters apply to both your prompt and the generated video, and enforcement is region-dependent. Policy-violating prompts or content are blocked.
Not supported	Uploading audio references; chained editing (submit a fresh `video_url` for each edit); voice/dialogue editing; and video extension or frame interpolation (this operation transforms existing footage, it does not lengthen it or generate new frames).
Regional limits	Editing an uploaded video is not available for users in the EEA, Switzerland, or the UK.

Prompting Guide

The prompt is passed through to the model verbatim, so all of Google’s prompt syntax works as-is.

Technique	How to use it
Keep it simple	Short prompts edit best; long paragraphs cause unintended changes. Prefer “Make this video anime” or “Change the lighting to be more dramatic”.
“Keep everything else the same”	When editing one aspect, add this phrase to preserve the rest — e.g. “Add a cat that jumps onto his lap. Keep everything else the same.”
Single scene	By default the model may split into several shots. Force one continuous shot by saying so — “in a single continuous shot”, “no scene cuts”.
Remove elements	Use simple negatives in the prompt — “no dialogue”, “no embellishments”, “no extra sound effects”.
Prompt the audio	Describe the sound you want — “calm background music”, “a high-energy techno beat”, “a low tinny radio in the background”.
Timing events	Use natural language (“after 3 seconds a woman enters”) or timecode syntax: `[0-5s] walking [5-10s] stops and turns [10-15s] starts running`.
On-screen text	Quote it exactly — e.g. a street sign that says: “This is an AI generation”.
Meta-prompting	Ask for general qualities — “be extremely detailed in your descriptions of characters and environments”.

Gemini Omni 1.2 Flash API Pricing

Resolution	Price (USD)
Per second of output video	$0.07297

⚡ Performance

Live usage measured on Pixazo's gateway, split by model version. Generation time is how long a generation takes end-to-end (lower is better). Success rate is the percent of generations that complete (higher is better).

Show data for the last

Generations

2,700last 30d

~90 per day

Success rate

100.0%

of completed generations

Generation time

2.0minavg

p95 2.1min

Requests

Oct 16max 1,500Nov 14

Gemini Omni 1.2 FlashAvg 90/day

Generation Time

Oct 16max 2.1minNov 14

Gemini Omni 1.2 FlashAvg 2.0min

Error Rate

Oct 16max 5.0%Nov 14

Gemini Omni 1.2 FlashAvg 0.0%

〰 Uptime

Percent of generations that succeeded over the selected period, per model version.

Avg. Success Rate (30d)

100.00%

across all generations of this model family

Uptime

Oct 16max 100%Nov 14

SuccessfulAvg 100.00%