Activity
Mon
Wed
Fri
Sun
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
What is this?
Less
More
Owned by Wei
A community for premium members. We share great ideas and battle-tested ComfyUI workflows for AI portrait editing.
A community for ComfyUI creators! Share workflows, swap ideas, and collaborate to level up your AI art game! MyAIForce+: https://myaiforce.com/pg
Memberships
26 contributions to MyAIForce
2d •
Best Way to Use Qwen Image Edit: My Secret Multi-Image Workflow Revealed!
YouTube: https://youtu.be/cU75sP6o3fg Qwen Edit 2509 surprised me again—here’s a fast, flexible ComfyUI workflow for multi-image editing. Watch me blend products into splash shots, drop portraits into new scenes, do clothing/accessory swaps, and keep backgrounds + poses consistent with prompt-controlled inpainting. We’ll cover FastCanvas positioning (lock/reset), defining masks with Qwen Image Edit, Inpaint Crop, expanding masks (Mask Box Extend), and a quick second repaint pass for sharper faces. Join our paid group and try it free for 7 days to get the workflow: https://myaiforce.com/pg
0 likes • 5h
@Robert Zavadsky It might be due to the input image being too large. I've also run into errors when uploading really big images, but the issues I face are a bit different from yours.
0 likes • 4h
I've conducted numerous tests and didn't notice any significant differences between the inpainting area and the surrounding areas. Could you share a full screenshot of your workflow? As for FastCanvas, I haven't experienced the same issues. Is your ComfyUI front end updated to the latest version? If you're not fond of this node, you can always manually paint the inpainting mask—check out the attached image for reference!
11d •
Beat Qwen Edit 2509 with This Flux Kontext Workflow
YouTube: https://youtu.be/i1dU04mvBYo In this video, we explore the new features of Qwen Image Edit 2509, which now supports multiple input images and OpenPose skeletons for easier pose transfer—no LoRA needed. We also compare Qwen Edit with Flux Kontext, highlighting its strengths and weaknesses. Despite Qwen’s improvements, Flux Kontext often outperforms it, especially in challenging scenarios. I also walk you through a detailed Flux Kontext workflow using ControlNet, offering tips to achieve better image quality and pose accuracy. Join our paid group and try it free for 7 days to get the workflow: https://myaiforce.com/pg
2
0
14d •
Qwen-Image-Edit-2509 vs. ACE++ for Clothes Swap
I use these different techniques for clothes swapping; which one do you think is better? I’ve shared some comparison images on Reddit(https://www.reddit.com/r/StableDiffusion/comments/1nquefe/qwenimageedit2509_vs_ace_for_clothes_swap/) to give us more options.
0 likes • 13d
@Robert Zavadsky Thanks for sharing your version! I noticed you included the lightning LoRA in your workflow, which can boost speed but may compromise image quality. I tried running without the lightning LoRA and took 20 sampling steps, but the image quality still fell short compared to Kontext. I suspect the training images for Qwen Editing might not be high enough in quality. Perhaps the full version of Qwen Edit could yield better results! Qwen Edit has stricter size limitations compared to Kontext. Qwen Edit should stay within 1K, while Kontext allows for up to 2K. I’ve noticed that if the output exceeds 1K with Qwen, a white band appears along the edge of the image.
16d •
Qwen-Image-Edit 2509 Shows Significant Improvement
Download the workflow below⏬ Qwen-Image-Edit has been updated to version 2509. It now supports 1-3 input images, with improved consistency. I developed a pose transfer workflow today. Previously, pose transfer with Qwen-Edit relied on LoRA, similar to the workflow I shared earlier (https://www.skool.com/myaiforce-5872/qwen-pose-transfer-workflow?p=c0f6c63e). The previous output was a stitched image of the two input images. With Qwen-Image-Edit 2509, the new workflow can generate images with new poses, eliminating the need to crop the output and reducing the image size.
20d •
Bye-Bye Plastic Skin! Real Skin Texture with SUPIR & SRPO Upscaling
YouTube: https://youtu.be/Q-9Wbk_AX7c In this video, I demonstrate a powerful workflow that enhances AI-generated skin, turning smooth, plastic-like surfaces into realistic, detailed textures. 💆♀️🔍 Watch as I compare before and after images, showcasing the workflow's impact on skin appearance. 🔑 Key Workflow Stages - Stage 1: Initial Upscaling (SUPIR) - Stage 2: Skin Enhancement (SRPO Fine-Tuned Model) Join our paid group and try it free for 7 days to get the workflow: https://myaiforce.com/pg
1-10 of 26
Active 4h ago
Joined Aug 9, 2025