Actions2
Create Task
AI-generatedSummary
Create an image generation task via the Comandos API.
Inputs
- model (required) — Model to use for generation. Options include Seedream 4.5, Flux Pro, GPT Image 1.5, Nano Banana, Nano Banana Pro, Nano Banana 2, Midjourney. Default: seedream. Note: Prompt is required per source implementation; throws NodeOperationError if empty.
- prompt — Описание изображения (макс. 3000 символов). Default: empty string. Source enforces non-empty; will throw if missing.
- ratio — Aspect ratio for the output. Model-dependent visibility: hidden for gpt-image-1.5 and nano-banana-2; separate optionsets for nano-banana-2 (includes 'auto') and gpt-image-1.5 (1:1, 2:3, 3:2). Default: 2:3.
- quality — Quality setting. Two UI variants: for seedream/flux-pro offers basic/high/medium (default basic); for gpt-image-1.5 offers medium/high (default medium). Other models default to basic quality internally.
- resolution — Output resolution. Shown for flux-pro, nano-banana-pro, nano-banana-2. Options: 1K, 2K, 4K. Default: 1K.
- googleSearch — Boolean flag to enable web-grounding via Google Search. Only visible for nano-banana-2 model. Default: false.
- references — Fixed collection of reference URLs for Image-to-Image workflows. Each entry has a 'url' string field. Multiple values allowed. Shown for operation 'create'.
Output shape
array
Each item contains a 'json' object with the full API response merged with the generation payload's URL, the node version, and a '_comandosMediaLoop' object holding the original prompt and extracted references. Input binary data is passed through if present.
Examples
Example 1: Basic text-to-image generation with Seedream 4.5
Set model to 'seedream', provide a prompt string, optionally adjust ratio/quality/resolution. The node will POST a task to the Comandos API and return the task ID and generation metadata in the output JSON.