Generate Image - Markifact

Create brand-new images such as ads, banners, infographics, slides or creative visuals using AI. Supports multiple AI image models, including the leading Gemini Nano Banana model. For modifying an existing image, use the ai_edit_image operation instead.


App	AI
Operation ID	`ai_generate_image`
Type	Action
Connection	None
Credits per run	20
Agent / MCP tool	Yes
MCP	Not exposed via MCP

Inputs

Field	Type	Required	Default	Description
`prompt`	string	Yes	-	Description of the image you want to generate. This is a multi-line text area where you can: - Describe the subject, style, and composition - Include details about colors, lighting, and mood - Reference dynamic data for personalized content using {{nodeId__data}} syntax - Use line breaks to structure detailed descriptions Example: “A professional headshot of {{customerNodeId__data.name}} Corporate style, good lighting, blue background High quality, business professional look”
`model`	string or SelectableOption	Yes	-	The AI model to use for image analysis. Available options: ‘gemini-3.1-flash-image-preview’ - This is the RECOMMENDED model. AKA ‘Nano Banana 2’, You must use this model by default for best results, unless user request something else. ‘gemini-3-pro-image-preview’ - AKA ‘Nano Banana Pro’, ‘gpt-image-2’ - OpenAI’s most advanced image generation model, best for complex prompts and highest quality ‘gpt-image-1.5’ - OpenAI’s old model for image generation
`numImages`	integer	No	`1`	Number of image variations to generate (1-10). More variations give you more options to choose from.
`output_format`	string or SelectableOption	No	-	Format of the generated images (GPT-Image only). Available options: ‘png’ - High quality, supports transparency ‘jpg’ - Smaller file size, good for photos ‘webp’ - Modern format, good compression
`quality`	string or SelectableOption	No	-	Quality of the generated images (GPT-Image only).
`size`	string or SelectableOption	No	-	Size of the generated images (GPT-Image only). Available options: ‘1024x1024’ - Square ‘1536x1024’ - Landscape ‘1024x1536’ - Portrait
`background`	string or SelectableOption	No	-	Background type for the generated images (GPT-Image only). Available options: ‘transparent’ - No background (PNG only) ‘opaque’ - Solid background ‘auto’ - Let the model decide
`aspect_ratio`	string or SelectableOption	No	-	Aspect ratio for the generated images (Gemini only). Available options: ‘1:1’ - Square ‘1:4’ - Ultra Tall ‘1:8’ - Ultra Tall ‘2:3’ - Portrait ‘3:2’ - Portrait ‘3:4’ - Portrait ‘4:1’ - Ultra Wide ‘4:3’ - Portrait ‘4:5’ - Portrait ‘5:4’ - Portrait ‘8:1’ - Panorama ‘9:16’ - Portrait ‘16:9’ - Landscape ‘21:9’ - Wide Landscape
`resolution`	string or SelectableOption	No	-	Resolution for generated images (Gemini only). Available options: ‘512px’ ‘1K’ # Default ‘2K’ ‘4K’

SelectableOption

Field	Type	Required	Default	Description
`value`	string	Yes	-
`label`	string	Yes	-

Output

Type: str | List[str] Returns URL(s) of generated image(s). Single URL if one image, list if multiple.