Omni Flash API with APIDot

Build with the Omni Flash API using APIDot: cURL, Node.js, request shapes, polling, webhooks, and production notes.

Try on APIDot | Get API Key | API Docs | Pricing | Main Examples

What this example shows

• How to submit Omni Flash multimodal video jobs through APIDot.
• How to choose text-only, image-guided, or one-video-guided request shapes.
• How to avoid invalid media combinations, including two image inputs or duration with video input.
• How to store data.task_id, poll status, and use webhook callbacks.

When to use it

Use this repo when a video workflow may start from a prompt, one image, three images, or one source video. It is useful for multimodal product previews, short-video revisions, visual direction tests, and backend queues that need one APIDot async workflow across several input shapes.

Requirements

• An APIDot account.
• An APIDot API key stored server-side.
• Node.js 18 or newer for the Node.js example.
• curl for the cURL example.
• Public image or video URLs when using guided workflows.

Environment variables

Use placeholders only. Do not commit real credentials.

APIDOT_API_KEY=YOUR_API_KEY_HERE
APIDOT_BASE_URL=https://api.apidot.ai
# Optional: set only when you have a real public webhook receiver.
# APIDOT_CALLBACK_URL=https://example.com/api/apidot/webhook

How to run

cp .env.example .env
# Edit .env and set APIDOT_API_KEY
cd node
npm start

Use curl/generate.md when you want a copy-paste cURL request instead of the Node.js polling example.

Minimal API request

Submit to APIDot's unified async generation endpoint:

{
  "model": "omni-flash",
  "callback_url": "https://example.com/api/apidot/webhook",
  "input": {
    "prompt": "A cinematic product reveal of a translucent speaker on wet black stone, slow push-in, realistic reflections, soft morning light.",
    "resolution": "720p",
    "duration": 6,
    "aspect_ratio": "16:9"
  }
}

Request variants

Text-only video

{
  "model": "omni-flash",
  "input": {
    "prompt": "A clean product reveal in a bright studio with a slow push-in and realistic reflections.",
    "resolution": "720p",
    "duration": 6,
    "aspect_ratio": "16:9"
  }
}

One-image-guided video

{
  "model": "omni-flash",
  "input": {
    "prompt": "Use the product image as the visual anchor and create a short premium motion clip.",
    "image_urls": [
      "https://example.com/product-reference.webp"
    ],
    "resolution": "1080p",
    "duration": 6,
    "aspect_ratio": "16:9"
  }
}

Three-image visual direction

{
  "model": "omni-flash",
  "input": {
    "prompt": "Create a short launch clip using these three images as product, scene, and style references.",
    "image_urls": [
      "https://example.com/product.webp",
      "https://example.com/scene.webp",
      "https://example.com/style.webp"
    ],
    "resolution": "1080p",
    "duration": 8,
    "aspect_ratio": "9:16"
  }
}

One-video-guided revision

{
  "model": "omni-flash",
  "input": {
    "prompt": "Revise the source clip with a cleaner product finish and brighter studio lighting.",
    "video_urls": [
      "https://example.com/source-video.mp4"
    ],
    "resolution": "1080p",
    "aspect_ratio": "16:9"
  }
}

Request parameters

Field	Type	Required	Notes
model	string	yes	Use omni-flash.
callback_url	string	no	Optional webhook callback URL for terminal task updates.
input.prompt	string	yes	Scene, motion, camera, and revision instruction.
input.image_urls	string[]	no	Send zero, one, or three public image URLs.
input.video_urls	string[]	no	Send at most one public video URL.
input.resolution	string	no	Supported values are 720p, 1080p, and 4k.
input.duration	integer	no	Supported values are 4, 6, 8, and 10 for jobs without input.video_urls.
input.aspect_ratio	string	no	Supported values are 16:9 and 9:16.

Expected response

Submit response:

{
  "code": 200,
  "data": {
    "task_id": "task-unified-example",
    "status": "not_started",
    "created_time": "2026-04-19T21:19:42"
  }
}

Shortened finished response:

{
  "code": 200,
  "data": {
    "task_id": "task-unified-example",
    "status": "finished",
    "output": {
      "files": [
        {
          "file_url": "https://example.com/generated-video.mp4",
          "file_type": "video"
        }
      ]
    }
  }
}

Production notes

• Keep APIDot API keys in server-side environment variables.
• Persist task_id, request shape, media URLs, selected resolution, user ID, and status together.
• Validate image count before submission; two images are not accepted by the current request contract.
• Omit input.duration whenever input.video_urls is present.
• Use polling for local tests and webhook callbacks for production queues.
• Avoid logging API keys, private prompts, private media URLs, or production callback URLs.

Common mistakes

• Sending two image URLs.
• Sending more than one video URL.
• Sending duration with video_urls.
• Assuming unsupported aspect ratios are accepted.
• Losing data.task_id before polling or webhook reconciliation.
• Committing a real .env file or API key.

Related links

• Website: https://apidot.ai
• Docs: https://apidot.ai/docs
• GitHub: https://github.com/APIDotAI
• Omni Flash model page: https://apidot.ai/models/omni-flash
• Omni Flash API docs: https://apidot.ai/docs/omni-flash
• APIDot quickstart: https://apidot.ai/docs/quickstart
• APIDot webhooks: https://apidot.ai/docs/webhooks
• Main APIDot examples repo: https://github.com/APIDotAI/apidot-examples

Promotion

• Topic: Omni Flash multimodal video request-shape selection and async APIDot workflow.
• Audience: Backend developers adding prompt, image-guided, or source-video-guided generation to product workflows.
• Tweet angle: Choose the Omni Flash input shape first: prompt only, one image, three images, or one video, then wire polling or webhooks.
• Landing page: https://apidot.ai/models/omni-flash
• Source status: verified

Quality review

• Quality score: 38/40
• Promotion candidate: yes
• Blocking issues: None for request-shape and async workflow promotion. Verify the model page and docs again before making cost, ranking, or availability claims.