comfyui-qwen-image-edit-node

ComfyUI Wuli Qwen Image Edit 2.0 Node

A custom ComfyUI node for:

• Text to image (TXT_2_IMG)
• Image to image (REF_2_IMG) with up to 4 reference images

It is implemented against the same Wuli open-platform API flow used by the wuli skill:

1. Upload reference images to Wuli OSS
2. Submit generation task
3. Poll task status
4. Fetch no-watermark URLs when available
5. Download outputs back into ComfyUI IMAGE

Install

1. Copy this folder into your ComfyUI custom nodes directory:

ComfyUI/custom_nodes/comfyui_qwen_image_edit

2. Restart ComfyUI.

Authentication

You can authenticate in either way:

• Fill the node input api_token
• Or set env var WULI_API_TOKEN

Windows PowerShell example:

$env:WULI_API_TOKEN = "wuli-your-token"

Node Inputs

• prompt: positive prompt
• negative_prompt: optional negative prompt
• mode:
  • auto: no refs -> text2img, has refs -> img2img
  • txt2img: force text2img
  • img2img: force img2img (requires at least 1 ref)
• model_name: Qwen Image 2.0 / Qwen Image Turbo
• aspect_ratio: output ratio
• resolution: 2K / 4K
• n: output count (1-4)
• optimize_prompt: API-side prompt optimization
• seed: -1 for random
• poll_interval: status polling interval (seconds)
• timeout_seconds: overall timeout
• ref_image_1 ... ref_image_4: optional reference images

Outputs

• images (IMAGE): generated image batch
• urls (STRING): one URL per line
• record_id (STRING): Wuli task record id

Notes

• This node uses remote API calls and may take time to complete.
• If no-watermark URL retrieval fails for a result, it falls back to the default result URL.