ComfyUI Topaz Upscaler (API)

A custom node that integrates the Topaz Labs Image API. This allows you to use Topaz's industry-leading AI models directly within your ComfyUI workflows.

Features

• Full Suite Support: Access Enhance, Sharpen, Denoise, Restore, and Lighting modes.
• Model Selection: Choose specific models like Standard V2, CGI, Lens Blur, etc.
• Batch Support: Automatically processes image batches sequentially.
• Auto-Scaling: Use a simple multiplier (e.g., 2.0x, 4.0x) or define specific output resolutions.
• Secure: Supports Environment Variables for API keys.

Installation

1. Clone this repository into your ComfyUI custom_nodes folder:

   cd ComfyUI/custom_nodes
   git clone [https://github.com/yourusername/comfyui-topaz-upscaler.git](https://github.com/yourusername/comfyui-topaz-upscaler.git)
   

2. Install the required Python package:

   pip install requests
   

   (Note: torch, numpy, and Pillow are likely already installed in your ComfyUI environment).

Configuration (API Key)

You need a generic API Key from Topaz Labs. You can configure this in two ways:

Option A: Environment Variable (Recommended)

Set the TOPAZ_API_KEY environment variable on your system.

• Windows (PowerShell): $env:TOPAZ_API_KEY="your_key_here"
• Linux/Mac: export TOPAZ_API_KEY="your_key_here"

If this is set, you can leave the api_key field in the node empty.

Option B: Node Input

Paste your API key directly into the api_key widget on the node itself. This overrides the environment variable.

Usage

1. Add the node via Image -> Topaz -> Topaz Upscaler (API).
2. Connect an image to the input.
3. Select your Mode (e.g., enhance for upscaling, denoise for cleaning).
4. Select the Model relevant to that mode (e.g., Standard V2 for enhance).
5. Upscaling:
   • Set scale_multiplier > 1.0 (e.g., 2.0) to automatically upscale.
   • Set scale_multiplier to 1.0 to manually control output_width and output_height.
6. Adjust strengths (denoise_strength, sharpen_strength) as needed.

Parameters

Parameter	Description
scale_multiplier	Multiplies input dimensions. Set to 1.0 to disable auto-scaling.
mode	Operation type (Enhance, Sharpen, Denoise, Restore, Lighting).
model	Specific AI model to use. Ensure the model matches the mode.
face_enhancement	Attempts to restore faces in the image (Enhance mode only).
output_format	Format sent to Topaz and received back (jpeg/png/tiff).

Requirements

• ComfyUI
• Internet connection (images are processed on Topaz servers)
• Topaz Labs API Credits