ComfyUI > Nodes > ComfyUI-KEEP

ComfyUI Extension: ComfyUI-KEEP

Repo Name

ComfyUI-KEEP

Author
wildminder (Account age: 4890 days)
Nodes
View all nodes(3)
Latest Updated
2025-07-04
Github Stars
0.07K

How to Install ComfyUI-KEEP

Install this extension via the ComfyUI Manager by searching for ComfyUI-KEEP
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-KEEP in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

ComfyUI-KEEP Description

ComfyUI-KEEP enhances video face super-resolution by using Kalman-inspired feature propagation, improving clarity and detail in facial features across video frames.

ComfyUI-KEEP Introduction

ComfyUI-KEEP is an innovative extension designed to enhance the quality of faces in images and video sequences using advanced super-resolution techniques. This extension integrates the Kalman-Inspired Feature Propagation (KEEP) algorithm into the ComfyUI framework, allowing AI artists to achieve high-quality face restoration effortlessly. Whether you're working with single images or entire video sequences, ComfyUI-KEEP provides robust tools to improve facial details, making it an invaluable asset for artists looking to elevate their digital creations.

How ComfyUI-KEEP Works

At its core, ComfyUI-KEEP leverages the KEEP algorithm, which is inspired by the Kalman filter—a mathematical approach used to predict and correct errors in dynamic systems. In the context of video face super-resolution, KEEP uses this principle to propagate and enhance facial features across frames, ensuring consistent and high-quality results. Imagine it as a smart tool that not only sharpens and clarifies faces but also maintains their natural appearance across multiple frames, much like how a skilled artist might refine a portrait over time.

ComfyUI-KEEP Features

ComfyUI-KEEP offers a range of features tailored to meet the needs of AI artists:

  • High-Quality Face Restoration: Utilizes the KEEP model to enhance facial details, ensuring that faces appear clear and lifelike.
  • Flexible Upscaling Integration: Connect any ComfyUI-compatible upscale model to enhance both the background and faces. This flexibility allows you to use different models for different parts of the image, providing full control over the final output.
  • Single Image & Sequence Processing: Whether you're working with a single image or a sequence of video frames, ComfyUI-KEEP can handle both, making it versatile for various projects.
  • Choice of Detection Models: Supports multiple face detection models, such as RetinaFace and YOLOv5, allowing you to choose the best fit for your project.

ComfyUI-KEEP Models

ComfyUI-KEEP includes several models to cater to different needs:

  • KEEP Model: The general model suitable for most face enhancement tasks.
  • Asian Model: A specialized model designed to enhance faces with Asian features, ensuring cultural and ethnic diversity in your projects. These models can be selected based on the specific requirements of your project, allowing for tailored enhancements that respect the unique characteristics of different faces.

What's New with ComfyUI-KEEP

Recent updates have introduced several enhancements to ComfyUI-KEEP:

  • 2025-07-04: Added multiface processing and tracking, allowing for the enhancement of multiple faces within a single frame.
  • 2025-07-03: Introduced standard ComfyUI Upscale Model inputs, replacing the previously hardcoded Real-ESRGAN, and added a final_upscale_factor for precise output scaling. These updates improve the flexibility and precision of the extension, making it easier for artists to achieve their desired results.

Troubleshooting ComfyUI-KEEP

Here are some common issues you might encounter while using ComfyUI-KEEP, along with solutions:

  • Issue: The extension is not enhancing faces as expected.
  • Solution: Ensure that the correct model is selected in the Load KEEP Models node. Verify that the input images are of sufficient quality and that the face detection model is appropriate for the task.
  • Issue: The output size is not as desired.
  • Solution: Adjust the final_upscale_factor to control the final output size. For example, if your background upscale model is set to 4x but you want a 2x final output, set the factor to 2.0.
  • Issue: Multiple faces are not being processed.
  • Solution: Ensure that multiface processing is enabled and that the detection model supports multiple face tracking.

Learn More about ComfyUI-KEEP

To further explore the capabilities of ComfyUI-KEEP, consider visiting the following resources:

  • Original KEEP Project: Learn more about the underlying KEEP algorithm and its applications.
  • Demo Video: Watch a demonstration of the KEEP algorithm in action.
  • Community Forums: Join discussions with other users and developers to share insights and seek support. These resources provide valuable insights and support, helping you make the most of ComfyUI-KEEP in your artistic endeavors.

ComfyUI-KEEP Related Nodes

RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

ComfyUI-KEEP detailed guide | ComfyUI