Video generation

The Otoroshi LLM extension provides video generation capabilities through an OpenAI-compatible API. You can configure video generation providers, expose them through Otoroshi routes, and let consumers generate videos from text prompts.

Supported providers

Provider	Text-to-Video	Default model
Luma	Yes	ray-flash-2

Currently, Luma (Dream Machine API) is the only supported video generation provider.

Features

• Text-to-video generation from text prompts
• Model routing — route to different providers using provider/model syntax
• Model constraints — restrict which models consumers can use via include/exclude regex patterns, enforceable per API key or per user
• Budget enforcement — video generation costs are tracked and budgets are enforced
• Cost tracking — per-request cost tracking integrated with cost tracking
• Configurable parameters — control aspect ratio, resolution, duration, and looping

API endpoint

The extension exposes a video generation endpoint through an Otoroshi route plugin:

Endpoint	Method	Description
/v1/videos/generations	POST	Generate a video from a text prompt

Plugin configuration

Add the Cloud APIM - Video generation backend plugin to your route:

{
  "enabled": true,
  "plugin": "cp:otoroshi_plugins.com.cloud.apim.otoroshi.extensions.aigateway.plugins.VideosGen",
  "config": {
    "refs": ["video-model-entity-id"]
  }
}

Parameter	Type	Default	Description
refs	array	[]	List of Video Model entity IDs

Generation request

curl --request POST \
  --url http://myroute.oto.tools:8080/v1/videos/generations \
  --header 'content-type: application/json' \
  --data '{
  "prompt": "A cat playing with a ball of yarn in a sunny garden",
  "model": "ray-flash-2",
  "aspect_ratio": "16:9",
  "resolution": "720p",
  "duration": "5s",
  "loop": false
}'

Generation parameters

Parameter	Type	Description
prompt	string	The text prompt describing the video to generate
model	string	Model name (can include provider prefix)
aspect_ratio	string	Aspect ratio (e.g., 16:9)
resolution	string	Video resolution (e.g., 720p)
duration	string	Video duration (e.g., 5s)
loop	boolean	Whether the video should loop

Generation response

{
  "created": 1762441013,
  "data": [
    {
      "url": "https://...",
      "b64_json": null,
      "revised_prompt": null
    }
  ],
  "usage": {
    "total_tokens": 0,
    "input_tokens": 0,
    "output_tokens": 0
  }
}

Model routing

When multiple video model providers are configured in refs, you can target a specific provider using the model field in your request:

{
  "prompt": "a sunset over the ocean",
  "model": "providerName/modelName"
}

The provider can be referenced by:

• Entity name (slug): my-luma-provider/ray-flash-2
• Entity ID: video-model-id###ray-flash-2

If no provider prefix is specified, the first configured ref is used.

Entity configuration

A video model entity is configured with a provider type, connection settings, and generation options:

{
  "id": "video-model-entity-id",
  "name": "My Luma Provider",
  "description": "A Luma video generation provider",
  "provider": "luma",
  "config": {
    "connection": {
      "token": "${vault://local/LUMA_API_TOKEN}",
      "timeout": 180000
    },
    "options": {
      "model": "ray-flash-2",
      "aspect_ratio": "16:9",
      "resolution": "720p",
      "duration": "5s"
    }
  },
  "models": {
    "include": [],
    "exclude": []
  }
}

The token field supports comma-separated round-robin tokens — if the token contains commas, the system rotates through them.

Luma options

Parameter	Type	Default	Description
enabled	boolean	true	Enable or disable video generation
model	string	ray-flash-2	The Luma model to use
aspect_ratio	string	16:9	Video aspect ratio
resolution	string	720p	Video resolution
duration	string	5s	Video duration
loop	boolean	—	Whether the video should loop

Request-level parameters override entity-level configuration.

Luma models

Model	Description
ray-flash-2	Fast generation model
photon-1	High quality model