ComfyUI Upsampler Nodes

ComfyUI custom nodes for integrating with the Upsampler API to enhance and upscale images using AI.

Features

Three different upscaling methods:

1. Smart Upscale - Enhances images by intelligently regenerating at a higher resolution
2. Dynamic Upscale - Preserves the overall composition while forcing new details
3. Precise Upscale - Upscales and sharpens while completely preserving the original structure

Installation

1. Clone or download this repository to your ComfyUI custom_nodes directory:

   cd ComfyUI/custom_nodes
   git clone <this-repo-url> comfyui-upsampler
   

2. Install dependencies:

   pip install -r comfyui-upsampler/requirements.txt
   

3. Restart ComfyUI

Setup Required

1. Get Upsampler API Key

1. Visit https://upsampler.com/
2. Create an account and generate an API key
3. Add credits to your account for processing

2. Implement Image Hosting

IMPORTANT: Before using these nodes, you must implement the _upload_image_temp() method in nodes.py to upload images to a publicly accessible URL.

The Upsampler API requires images to be accessible via public URLs. You have several options:

Option A: Use ImgBB (Recommended)

1. Get a free API key from https://api.imgbb.com/
2. Uncomment and modify the ImgBB implementation in _upload_image_temp()

Option B: Use Your Own Server

Upload images to your own web server or CDN that provides public URLs.

Option C: Use Other Services

Implement integration with services like Imgur, Cloudinary, or AWS S3.

Usage

1. Add one of the Upsampler nodes to your ComfyUI workflow:

   • 🔍 Upsampler Smart Upscale
   • ⚡ Upsampler Dynamic Upscale
   • 🎯 Upsampler Precise Upscale

2. Connect an IMAGE input to the node

3. Configure the parameters:

   • API Key: Your Upsampler API key
   • Input Image Type: Choose "realism", "anime", or "universal"
   • Upscale Factor: How much to upscale (1.0-4.0 for Smart/Dynamic, 1.0-16.0 for Precise)
   • Additional parameters specific to each method

4. The node will:

   • Upload your image to a public URL
   • Submit the upscaling job to Upsampler API
   • Poll for completion with progress updates
   • Download and return the upscaled image

Node Parameters

Smart Upscale

• global_creativity (0-10): How much the AI can change the original image
• detail (0-10): Level of detail enhancement
• description: Text prompt to guide the upscaling process
• should_enhance_faces: Better preserve and enhance facial features
• should_preserve_blur: Preserve existing blur in the image

Dynamic Upscale

• global_creativity (0-10): How much the AI can change the original image
• resemblance (0-10): How closely to adhere to original image structure
• detail (0-10): Level of detail enhancement
• description: Text prompt to guide the upscaling process
• should_enhance_faces: Better preserve and enhance facial features
• should_preserve_hands: Better preserve hand structures
• should_preserve_blur: Preserve existing blur in the image

Precise Upscale

• should_enhance_faces: Apply face restoration techniques
• should_preserve_blur: Preserve existing blur in the image

Pricing

Credits are consumed based on output image size:

• Smart/Dynamic Upscale: 1 credit per megapixel (minimum 1 credit)
• Precise Upscale: 1 credit per 4 megapixels (minimum 1 credit)

Troubleshooting

1. "Image hosting not implemented" error: You need to implement the _upload_image_temp() method
2. API authentication errors: Check your API key is correct and has credits
3. Rate limit errors: The API has limits of 60 requests/minute and 1800 requests/hour
4. Job failures: Check the error message - common issues include invalid parameters or insufficient credits

API Documentation

For full API documentation, visit: https://upsampler.com/api

License

This custom node implementation is provided as-is. Please refer to Upsampler's terms of service for API usage.