Skip to main content

AZURE_OPENAI_IMAGE_GENERATION

ImageGeneration using Azure OpenAI.

Documentation Inheritance

Documentation partly inherited from OPENAI_IMAGE_GENERATION.

At a glance

  • Category LLM
  • Aliases SERVICE_AZURE_OPENAI_IMAGE_GENERATION
  • Version: 1.0.0
  • Applications: all
  • Scope: all
  • Default Service: AZURE_OPENAI_IMAGE_GENERATION

Config Options

NameDescriptionDefaultRequiredResolvedConstraintsConditional Rules
modelImage model to use (gpt-image-1, dall-e-3, dall-e-2)dall-e-3falsefalseNoneNone
instructionThe user prompt to send to the LLM (supports templates)NonefalsetrueLength:1-∞None
sizeDesired image size (model-dependent)NonefalsefalseNoneNone
qualityImage quality settingNonefalsefalseNoneNone
nNumber of images to generate (1-4; DALL·E3 supports 1)1falsefalseRange:1-4None
styleImage style for DALL-E 3 (natural, vivid)NonefalsefalseNoneConflicts: output_format
response_formatDALL·E response format: 'url' or 'b64_json' (not supported for gpt-image-1)NonefalsefalseNoneConflicts: output_format, background, seed
output_formatOutput format for gpt-image-1 only (png|jpeg|webp)NonefalsefalseNoneConflicts: response_format, style
backgroundBackground setting for gpt-image-1 only (transparent)NonefalsefalseNoneConflicts: response_format
seedRandom seed for reproducible results (gpt-image-1 only)NonefalsefalseNoneConflicts: response_format
userUser identifier for attribution and policy monitoringNonefalsefalseNoneNone
timeoutMsRequest timeout in milliseconds60000falsefalseRange:1000-900000None
maxRetriesMaximum number of retries for recoverable errors2falsefalseRange:-∞-20None
outputKeyFlowContext key where images will be stored (default: imageList)imageListfalsefalsePattern: ^[a-zA-Z_][a-zA-Z0-9_]*$
Length:1-∞
None

Step-Level Validation Rules

Inputs

TypeDescriptionDefaultRequiredResolved
textPrompt text input. If the step reads text input, use a preceding SET step to provide it (e.g. - step: SET text: "..." ).Nonefalsefalse

Outputs

TypeDescriptionOptional
imageListArray of generated image items (each item contains url or b64_json) stored under the outputKey or default keyfalse
objectOptional metadata about the generation (model, size, attemptCount)true

Examples

Generate two images from a prompt

- step: SET
text: "A futuristic skyline at sunset, cinematic lighting"
- step: AZURE_OPENAI_IMAGE_GENERATION
n: 2
size: 1024x1024

See Also

General Resources: