ComfyUI  >  Nodes  >  PuLID_ComfyUI >  Apply PuLID

ComfyUI Node: Apply PuLID

Class Name


cubiq (Account age: 5009 days)
Latest Updated
Github Stars

How to Install PuLID_ComfyUI

Install this extension via the ComfyUI Manager by searching for  PuLID_ComfyUI
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter PuLID_ComfyUI 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
  • High-speed GPU machines
  • 200+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 50+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

Apply PuLID Description

Enhance AI-generated images with PuLID model for improved fidelity, style, and face analysis precision.

Apply PuLID:

The ApplyPulid node is designed to enhance your AI-generated images by applying the PuLID (Perceptual Loss for Image Denoising) model. This node integrates various advanced techniques to improve the fidelity and style of your images, making them more visually appealing and realistic. By leveraging models like EVA-CLIP and InsightFace, ApplyPulid can perform detailed face analysis and apply perceptual loss adjustments to your images. This node is particularly useful for artists looking to refine their AI-generated images with high precision and control over various parameters such as weight, noise, and projection methods.

Apply PuLID Input Parameters:


This parameter expects a pre-trained model that will be used as the base for applying the PuLID adjustments. The model serves as the foundation upon which the perceptual loss techniques will be applied.


This parameter requires the PuLID model, which contains the specific algorithms and data necessary for applying perceptual loss to the image. It is essential for the node to function correctly.


This parameter takes the EVA-CLIP model, which is used for advanced image processing and enhancement. EVA-CLIP helps in refining the image by leveraging its powerful visual recognition capabilities.


This parameter requires a face analysis model, such as InsightFace, to perform detailed facial recognition and adjustments. This is particularly useful for enhancing portraits and images with human faces.


This parameter is the input image that you want to enhance using the PuLID model. The image will undergo various transformations based on the other input parameters.


This parameter allows you to choose the method of enhancement: fidelity, style, or neutral. Each method offers a different approach to image enhancement, allowing you to customize the output based on your artistic needs.


This parameter controls the intensity of the PuLID adjustments. It accepts a float value with a default of 1.0, a minimum of -1.0, and a maximum of 5.0, with a step of 0.05. Higher values result in more pronounced adjustments.


This parameter defines the starting point for applying the PuLID adjustments, expressed as a float between 0.0 and 1.0, with a default of 0.0. It allows you to control when the adjustments begin during the image processing.


This parameter sets the endpoint for the PuLID adjustments, also expressed as a float between 0.0 and 1.0, with a default of 1.0. It determines when the adjustments should stop, giving you finer control over the process.

attn_mask (optional)

This optional parameter allows you to provide an attention mask, which can be used to focus the PuLID adjustments on specific areas of the image. This is useful for targeted enhancements.

Apply PuLID Output Parameters:


The output is the enhanced model, which has undergone the PuLID adjustments based on the input parameters. This model can be used for further image generation or refinement, providing a more visually appealing and realistic result.

Apply PuLID Usage Tips:

  • Experiment with different method settings (fidelity, style, neutral) to see which one best suits your artistic vision.
  • Adjust the weight parameter to control the intensity of the enhancements. Start with the default value and gradually increase or decrease it to achieve the desired effect.
  • Use the start_at and end_at parameters to fine-tune the timing of the adjustments, especially if you are working with animations or sequences of images.
  • If you are focusing on enhancing specific areas of an image, consider using the attn_mask parameter to apply targeted adjustments.

Apply PuLID Common Errors and Solutions:

"Model not found"

  • Explanation: This error occurs when the specified model is not available or cannot be loaded.
  • Solution: Ensure that the model file path is correct and that the model is properly loaded before running the node.

"Invalid parameter value"

  • Explanation: This error is triggered when one or more input parameters have values outside their acceptable ranges.
  • Solution: Double-check the values of all input parameters, ensuring they fall within the specified min and max limits.

"Face analysis model not loaded"

  • Explanation: This error occurs when the face analysis model is not properly loaded or initialized.
  • Solution: Verify that the face analysis model is correctly specified and loaded before applying the PuLID adjustments.

"EVA-CLIP model not found"

  • Explanation: This error indicates that the EVA-CLIP model is missing or not correctly loaded.
  • Solution: Make sure the EVA-CLIP model is available and properly loaded into the node before proceeding.

"Image input missing"

  • Explanation: This error occurs when the input image is not provided.
  • Solution: Ensure that you have specified an input image for the node to process.

Apply PuLID Related Nodes

Go back to the extension to check out more related nodes.

© Copyright 2024 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals.