Stable Diffusion XL v1.0 VAE Fix

Identifier for the type of task being performed

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

Image output type.

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

• `JPG`: Best for photorealistic images with smaller file sizes (no transparency).
• `PNG`: Lossless compression, supports high quality and transparency (alpha channel).
• `WEBP`: Modern format providing superior compression and transparency support.

Compression quality of the output. Higher values preserve quality but increase file size.

Specifies a webhook URL where JSON responses will be sent via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

Determines how the API delivers task results.

Specifies a URL where the generated content will be automatically uploaded using the HTTP PUT method. The raw binary data of the media file is sent directly as the request body. For secure uploads to cloud storage, use presigned URLs that include temporary authentication credentials.

Common use cases:

• Cloud storage: Upload directly to S3 buckets, Google Cloud Storage, or Azure Blob Storage using presigned URLs.
• CDN integration: Upload to content delivery networks for immediate distribution.

// S3 presigned URL for secure upload
https://your-bucket.s3.amazonaws.com/generated/content.mp4?X-Amz-Signature=abc123&X-Amz-Expires=3600

// Google Cloud Storage presigned URL
https://storage.googleapis.com/your-bucket/content.jpg?X-Goog-Signature=xyz789

// Custom storage endpoint
https://storage.example.com/uploads/generated-image.jpg

The content data will be sent as the request body to the specified URL when generation is complete.

Content safety checking configuration for image generation.

Time-to-live (TTL) in seconds for generated content. Only applies when outputType is URL.

Include task cost in the response.

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

Advanced caching mechanisms to speed up generation.

Image used as a starting point for the generation (UUID, URL, Data URI, or Base64).

Image used to specify which areas of the seed image should be edited (UUID, URL, Data URI, or Base64).

Identifier of the model to use for generation.

Text prompt describing elements to include in the generated output.

Prompt to guide what to exclude from generation. Ignored when guidance is disabled (CFGScale ≤ 1).

Width of the generated media in pixels.

Height of the generated media in pixels.

Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.

Total number of denoising steps. Higher values generally produce more detailed results but take longer.

Scheduler to use for the diffusion process.

Guidance scale representing how closely the output will resemble the prompt. Higher values produce results more aligned with the prompt.

Strength of the transformation. Lower values result in more influence from the original input.

Extra context pixels around the masked region during inpainting. The model zooms into the masked area with these additional pixels for better integration.

Number of layers to skip in the CLIP model.

VAE model identifier. Overrides the default VAE included with the base model.

Defines the syntax to be used for prompt weighting.

Prompt weighting allows you to adjust how strongly different parts of your prompt influence the generated image. Choose between compel notation with advanced weighting operations or sdEmbeds for simple emphasis adjustments.

Pixel extensions for each boundary direction of the source image. At least one direction is required.

With LoRA (Low-Rank Adaptation), you can adapt a model to specific styles or features by emphasizing particular aspects of the data. This technique enhances the quality and relevance of generated content and can be especially useful when the output needs to adhere to a specific artistic style or follow particular guidelines.

Multiple LoRA models can be used simultaneously to achieve different adaptation goals.

"lora": [
  {
    "model": "<lora-model-air>",
    "weight": 0.8
  }
]

With ControlNet, you can provide a guide image to help the model generate images that align with the desired structure. This guide image can be generated with our ControlNet preprocessing tool, extracting guidance information from an input image. The guide image can be in the form of an edge map, a pose, a depth estimation or any other type of control image that guides the generation process via the ControlNet model.

Multiple ControlNet models can be used at the same time to provide different types of guidance information to the model.

"controlNet": [
  {
    "model": "<controlnet-model-air>",
    "guideImage": "c64351d5-4c59-42f7-95e1-eace013eddab",
    "weight": 0.7,
    "startStep": 0,
    "endStep": 20,
    "controlMode": "controlnet"
  }
]

IP-Adapters enable image-prompted generation, allowing you to use reference images to guide the style and content of your generations. Multiple IP Adapters can be used simultaneously.

"ipAdapters": [
  {
    "model": "<ip-adapter-model-air>",
    "guideImages": ["c64351d5-4c59-42f7-95e1-eace013eddab"],
    "weight": 0.75
  },
  {
    "model": "<ip-adapter-model-air>",
    "guideImages": ["d7e8f9a0-2b5c-4e7f-a1d3-9c8b7a6e5d4f"],
    "weight": 0.5
  }
]

Refiner models help create higher quality image outputs by incorporating specialized models designed to enhance image details and overall coherence. This can be particularly useful when you need results with superior quality, photorealism, or specific aesthetic refinements. Note that refiner models are only SDXL based.

"refiner": {
  "model": "<refiner-model-air>",
  "startStep": 30
}

Embeddings (or Textual Inversion) can be used to add specific concepts or styles to your generations. Multiple embeddings can be used at the same time.

"embeddings": [
  {
    "model": "<embedding-model-air>"
  }
]

Configuration object for Ultralytics face enhancement during generation. This feature uses face detection and inpainting to improve facial details in the same generation step, without requiring post-processing.

Two-stage generation for improved resolution and detail. The model generates at a lower resolution first, then upscales and refines the result in a second pass. Can be enabled with true for default settings, or configured as an object for fine-grained control over the upscaling model, steps, and strength.

When using the object form, the model parameter is required. Available upscaling models:

"hiresFix": true

"hiresFix": {
  "model": "runware:esrgan@ultrasharp",
  "steps": 15,
  "strength": 0.6
}

PhotoMaker enables personalized image generation while preserving the identity of a subject from reference images. Provide one or more photos of a person, and the model will generate new images that maintain their facial characteristics, expression tendencies, and overall identity across different styles and compositions.

"photoMaker": {
  "images": ["59a2edc2-45e6-429f-be5f-7ded59b92046"],
  "strength": 20,
  "style": "Cinematic"
}

Configuration object for adding watermarks to generated videos. Watermarks can be applied using either text or image content with customizable positioning and appearance. You must provide either text or image content for the watermark, but not both.

"advancedFeatures": {
  "watermark": {
    "text": "© 2025 Company",
    "displayPosition": "bottom-right",
    "opacity": 0.6,
    "fontColor": "#ffffff",
    "bgColor": "#000000"
  }
}

"advancedFeatures": {
  "watermark": {
    "image": "c64351d5-4c59-42f7-95e1-eace013eddab",
    "displayPosition": "top-left",
    "opacity": 0.6
  }
}

"advancedFeatures": {
  "watermark": {
    "text": "PREVIEW",
    "tiled": true,
    "opacity": 0.4,
    "fontColor": "#cccccc"
  }
}