Generate and edit videos with Omni

Create a Omni video task

curl --request POST \
  --url https://gengen.farm/v1/contents/generations/tasks \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "gemini-omni-1.1-flash-preview",
  "prompt": "A cinematic ocean sunrise with gentle surf sounds."
}
'

import requests

url = "https://gengen.farm/v1/contents/generations/tasks"

payload = {
    "model": "gemini-omni-1.1-flash-preview",
    "prompt": "A cinematic ocean sunrise with gentle surf sounds."
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: 'gemini-omni-1.1-flash-preview',
    prompt: 'A cinematic ocean sunrise with gentle surf sounds.'
  })
};

fetch('https://gengen.farm/v1/contents/generations/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "id": "omni:12345678-1234-4234-8234-123456789abc",
  "model": "gemini-omni-1.1-flash-preview",
  "status": "queued",
  "outputs": {
    "videos": [],
    "images": []
  },
  "createdAt": "2026-10-09T09:58:00.000Z",
  "updatedAt": "2026-10-09T09:58:00.000Z"
}

POST

contents

generations

tasks

Create a Omni video task

curl --request POST \
  --url https://gengen.farm/v1/contents/generations/tasks \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "gemini-omni-1.1-flash-preview",
  "prompt": "A cinematic ocean sunrise with gentle surf sounds."
}
'

import requests

url = "https://gengen.farm/v1/contents/generations/tasks"

payload = {
    "model": "gemini-omni-1.1-flash-preview",
    "prompt": "A cinematic ocean sunrise with gentle surf sounds."
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: 'gemini-omni-1.1-flash-preview',
    prompt: 'A cinematic ocean sunrise with gentle surf sounds.'
  })
};

fetch('https://gengen.farm/v1/contents/generations/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "id": "omni:12345678-1234-4234-8234-123456789abc",
  "model": "gemini-omni-1.1-flash-preview",
  "status": "queued",
  "outputs": {
    "videos": [],
    "images": []
  },
  "createdAt": "2026-10-09T09:58:00.000Z",
  "updatedAt": "2026-10-09T09:58:00.000Z"
}

Use gemini-omni-1.1-flash-preview or gemini-omni-flash-preview under Google. Both are Preview models and return asynchronous video tasks with native audio.

Create and retrieve

curl https://gengen.farm/v1/contents/generations/tasks \
  -H 'Authorization: Bearer gengen_live_xxxxxxxxxxxxxxxx' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gemini-omni-1.1-flash-preview",
    "mode": "text_to_video",
    "prompt": "An ocean sunrise with gentle surf sounds.",
    "controls": {"duration": 5, "resolution": "720p", "ratio": "16:9"}
  }'

Save the returned id, including omni:, and poll the shared task endpoint with the same API key until succeeded or failed. Read outputs.videos on success. Each creation request creates a new task; polling never resubmits generation.

curl 'https://gengen.farm/v1/contents/generations/tasks/omni:YOUR_TASK_UUID' \
  -H 'Authorization: Bearer gengen_live_xxxxxxxxxxxxxxxx'

Modes

Mode	Assets	Omni Flash	Omni 1.1 Flash
`text_to_video`	None	Yes	Yes
`video_modify`	`sourceVideo`	Yes	Yes
`image_first_frame`	`firstFrameImage`	No	Yes
`image_first_last_frame`	`firstFrameImage`, `lastFrameImage`	No	Yes
`multimodal_reference`	`referenceImages`, `referenceVideos`	No	Yes
`video_extend`	`sourceVideo`	No	Yes

Mode is inferred when omitted; a source video defaults to editing. Set video_extend explicitly to extend. Do not combine frames, references and source-video input forms. Audio input and image output are unsupported.

{
  "model": "gemini-omni-1.1-flash-preview",
  "mode": "image_first_last_frame",
  "prompt": "Move smoothly from the first frame to the last frame.",
  "assets": {
    "firstFrameImage": "https://example.com/start.png",
    "lastFrameImage": "https://example.com/end.png"
  },
  "controls": {"duration": 5, "resolution": "1080p", "ratio": "16:9"}
}

Limits and controls

prompt: 1–10,000 characters.
duration: requested duration, integer 3–10 seconds; default 5. Actual output and billable token usage may differ.
resolution: Omni Flash supports 720p. Omni 1.1 adds 360p, 1080p and 4k.
ratio: 16:9 or 9:16; default 16:9.
One video per task. outputCount, if supplied, must be 1. Playground can submit several independent tasks.
Native audio is always enabled. generateAudio, if supplied, must be true.
Up to 10 reference images and 3 reference videos. Images accept JPEG/PNG/WebP over HTTPS (JPEG/PNG for GCS), up to 20 MiB each. Converted PNG must also fit this limit. Videos accept MP4 up to 64 MiB and 10 seconds each.

Use public/signed HTTPS URLs or account-authorized gs:// objects. GCS access requires both caller authorization and storage read permissions; contact support to register an external object or directory. Base64, local paths and asset:// inputs are unsupported. Media is validated before upstream generation; invalid media fails the task and releases its reservation.

Billing and outputs

Billing combines input tokens, text/reasoning tokens and video output tokens. See current prices. A conservative maximum reservation covers generation; final settlement uses actual returned token usage and releases the unused reservation. A requested duration alone does not determine the final price. Video links are temporary HTTPS URLs, retained for seven days after completion. Retrieve the task to renew a link before retention expires, or copy it to your own storage. Listing owned tasks is supported. This GENGEN endpoint does not expose cancellation, deletion, webhooks, streaming, or conversational interaction IDs. Invalid or unsupported parameters return HTTP 400 before reservation. Insufficient balance returns 402. A task whose upstream acceptance is uncertain remains pending for reconciliation; retrieve it rather than submitting a duplicate.

Authorizations

Authorization

string

header

required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

model

enum<string>

required

Available options:

gemini-omni-1.1-flash-preview,

gemini-omni-flash-preview

prompt

string

required

Required string length: 1 - 10000

mode

enum<string>

Inferred from assets when omitted; sourceVideo defaults to video_modify. Omni Flash supports text_to_video and video_modify; all modes are supported by Omni 1.1 Flash.

Available options:

text_to_video,

image_first_frame,

image_first_last_frame,

multimodal_reference,

video_extend,

video_modify

assets

object

Choose frames, reference images/videos, or one sourceVideo for editing/extension. Do not mix these input forms. Images: 20 MiB; MP4 videos: 64 MiB and at most 10 seconds each.

Show child attributes

controls

object

Show child attributes

providerOptions

object

Show child attributes

Response

Owned asynchronous video task

string

required

Example:

"omni:12345678-1234-4234-8234-123456789abc"

model

enum<string>

required

Available options:

gemini-omni-1.1-flash-preview,

gemini-omni-flash-preview

status

enum<string>

required

Available options:

queued,

running,

succeeded,

failed

outputs

object

required

Show child attributes

createdAt

string<date-time>

updatedAt

string<date-time>

failureReason

string

Reference

BytePlus

OpenAI

Kling

MiniMax

Happy Horse

Google

Flash VSR

Wan

Qwen

Generate and edit videos with Omni

Create and retrieve

Modes

Limits and controls

Billing and outputs

Authorizations

Body

Response

​Create and retrieve

​Modes

​Limits and controls

​Billing and outputs

Authorizations

Body

Response

Create and retrieve

Modes

Limits and controls

Billing and outputs