Aliyun Bailian | 阿里云百炼Install

Summary

Generate images from text prompts using various specialized and high-resolution text-to-image models provided by Aliyun Bailian.

Inputs

• imageModel (required) — Select the text-to-image model variant to use, each with different capabilities such as resolution limits, speed, style support, and text rendering quality.
• imagePrompt (required) — Primary text prompt describing the scene or subject to generate.
• imageNegativePrompt — Optional negative prompt to specify undesired elements to avoid in the generated image.
• imageSize — Select or specify the resolution of the generated image; constraints vary by model. For wan2.6-t2i model, resolution must be within specific size and aspect ratio rules.
• imageCount — Number of images to generate; some models limit this (e.g., Z-Image Turbo supports only 1 image).
• imageReferenceImage — Optional URL of a reference image used to guide repainting or style transfer for supported models.
• imageStyle — Preset output style to influence image aesthetics, supported only by certain models such as wanx-v1.
• imageAdvancedOptions — Advanced settings including prompt extension (auto prompt refinement), reference image usage mode (repaint or style transfer), reference strength, random seed for reproducibility, and watermark overlay.
• pollInterval — Polling delay in milliseconds while waiting for task completion if the node is set to wait for the task.
• maxWaitSeconds — Maximum wait time in seconds before timing out when polling for task completion.

Output shape

a JSON object containing URLs or data of the generated images corresponding to the requested count

The node submits a generation task to the Aliyun Bailian service and optionally polls until the task completes or times out. The output includes generated image data or URLs. Available image counts and resolutions depend on the selected model. Some models support reference image based editing and advanced options like style and watermark.

Examples

Example 1: Generate a single high-quality image of a cute orange cat basking in sunlight

Select model 'Wan 2.5 T2I Preview', enter a descriptive prompt, choose desired resolution, and optionally set a seed for reproducibility.

Example 2: Quickly generate one image with complex text rendering

Select 'Z-Image Turbo' model, provide prompt, generate 1 image (model constraint).

Links

• n8n community nodes installation guide