azure-flux-comfyui

Azure Flux ComfyUI Nodes

Minimal, production-focused ComfyUI nodes for Azure OpenAI Images (Flux). Provides:

• FLUX.1-Kontext-ImageToImage: edit existing image (edits API)
• FLUX.1-Kontext-TextToImage: generate from text (generations API)
• FLUX-1.1-pro-TextToImage: generate from text (generations API)
• FLUX.2-pro-TextToImage: generate from text (generations API)

All nodes call the official Azure OpenAI Images endpoints. No extra services, wrappers, or local models.

Overview

Four nodes with only the necessary fields for each mode. The API logic is unchanged and mirrors Azure’s Images Edits/Generations behavior.

Installation

Manual install:

cd ComfyUI/custom_nodes/
git clone https://github.com/xpanceo-team/azure-flux-comfyui.git
cd azure-flux-comfyui
pip install -r requirements.txt

Restart ComfyUI after installation.

Configuration

Set these fields in the node UI:

• azure_endpoint: https://<your-resource>.services.ai.azure.com/ or https://<your-resource>.cognitiveservices.azure.com/
• api_version: 2025-04-01-preview (or preview for FLUX.2-pro if required)
• api_key: your Azure OpenAI API key

Notes:

• Endpoint must end with /.
• Ensure the deployment name exists in your Azure resource:
  • Kontext nodes use FLUX.1-Kontext-pro
  • FLUX-1.1 node uses FLUX-1.1-pro
  • FLUX.2 node uses FLUX.2-pro

Nodes

• FLUX.1-Kontext-ImageToImage

  • Mode: image-to-image (edits)
  • Inputs: azure_endpoint, api_version, api_key, editing_prompt, image_1
  • Optional: size (WIDTHxHEIGHT, default 1024x1024), n (1–8), output_format (png|jpeg|webp)

• FLUX.1-Kontext-TextToImage

  • Mode: text-to-image (generations)
  • Inputs: azure_endpoint, api_version, api_key, editing_prompt
  • Optional: size, n, output_format

• FLUX-1.1-pro-TextToImage

  • Mode: text-to-image (generations)
  • Inputs: azure_endpoint, api_version, api_key, editing_prompt
  • Optional: size, n, output_format

• FLUX.2-pro-TextToImage

  • Mode: text-to-image (generations)
  • Inputs: azure_endpoint, api_version, api_key, editing_prompt
  • Optional: size, n, output_format

Parameters

• editing_prompt: text description or edit instruction
• image_1: input image tensor (only for ImageToImage)
• size: WIDTHxHEIGHT (e.g. 1024x1024)
• n: number of images to generate (1–8)
• output_format: png, jpeg, or webp

Requirements

• Python >= 3.8
• ComfyUI (current version)
• Dependencies (from requirements.txt):
  • requests
  • Pillow
  • torch
  • numpy

Troubleshooting

• 401/403: check api_key, resource/region, and endpoint URL.
• 404/400: verify the deployment name exists and supports the operation.
  • FLUX-1.1-pro supports generations; use TextToImage, not edits.
• Invalid size: use WIDTHxHEIGHT (e.g., 1024x1024).
• Image-to-image: image_1 is required.
• Timeouts: reduce n or try a smaller size.

License

MIT — see LICENSE.