ComfyUI: Transform 3D scenes with Qwen2-VL (Qwen-Edit)

Matt Hallett Visual·Jan 2026

Demo summary

The creator uses the Qwen-Edit workflow to perform complex scene modifications, such as changing weather to rain or snow, while preserving the original 3D geometry and text.

Step-by-step

Download the Qwen-Edit workflow from the website
Input the original 3D rendering or screenshot into the workflow
Enter a prompt specifying the desired time of day, season, and weather
Specify specific objects to add, such as cars or people with umbrellas
Adjust the noise settings to refine the image quality
Run the generation to apply the environmental changes while preserving geometry

Options

Use GGUF models to potentially reduce resource load
Manually flip car directions or regenerate with specific directional prompts (e.g., 'right to left')

Watch out for

The workflow is resource-heavy and may require high-end hardware like a 5090 GPU
Output quality can appear cartoonish or lack photorealism, especially regarding skin textures
AI-generated cars may face the wrong direction and require manual correction
People may be placed in unnatural scattering patterns

Tips

Be specific in prompts, such as 'people walking with umbrellas' rather than just 'add people'
Use Photoshop to paint over and fix minor errors like car orientation
Leverage the workflow to maintain specific geometry and text from the original scene

Highlights

“the workflow is really simple... it's just phenomenal. It's keeping the things you want and adding all the things you you would love to have.”

All demos from “Local "AI Rendering" Using 3D - Explained by a Human”

3:041:18Enhance 3D clay models with Flux and ControlNetThe creator demonstrates using the Flux model in ComfyUI with a Depth ControlNet to transform a gray 3D massing model into a detailed streetcar scene while maintaining structural integrity.ComfyUI· Image to Image
6:401:29Train a custom LoRA on CivitaiA step-by-step walkthrough of uploading a dataset of streetcar images to Civitai, using auto-labeling, and configuring training settings to create a custom LoRA model.Civitai· AI Image Generator
8:512:02Image-to-image enhancement with denoisingThe video shows how to use the 'denoising' strength in ComfyUI to blend a base 3D rendering with AI-generated details like autumn trees and realistic people.ComfyUI· Image to Image
12:211:24Text-to-image generation with Z-Turbo and QwenDemonstration of the Z-Turbo model and Qwen text encoder to generate specific advertising text on the side of a 3D tram model within a ComfyUI workflow.ComfyUI· Text to Image
15:063:12Transform 3D scenes with Qwen2-VL (Qwen-Edit)CurrentThe creator uses the Qwen-Edit workflow to perform complex scene modifications, such as changing weather to rain or snow, while preserving the original 3D geometry and text.ComfyUI· AI Inpainting
20:221:07Fix text and faces using Crop and StitchA demonstration of a custom 'crop image' node to isolate specific areas like signs or faces, regenerate them at native resolution, and stitch them back into the high-res image.ComfyUI· AI Inpainting
23:511:57Animate 3D renders with Wan 2.1The creator walks through a Wan 2.1 video generation workflow in ComfyUI, using the Painterly I2V node to control motion speed and animate a static 3D render.ComfyUI· Image to Video
Watch “Local "AI Rendering" Using 3D - Explained by a Human” →

AI Inpainting

16:571:43Generate composite image with Flux 2The creator demonstrates image-to-image generation by uploading two reference photos and using a text prompt to replace a logo on a coffee can with his own headshot using the Flux model.WINBUSH
20:010:28Text-to-image manipulation in ComfyUIThe video shows how to use a text prompt to attempt to rotate an object in an image within a specific node-based workflow.WINBUSH
1:111:34Configure LTX Latent Anchor Aware nodeThe video shows how to set up the LTX Latent Anchor Aware node in a ComfyUI workflow, explaining how to connect the reference image and adjust strength parameters to prevent identity drift.The AI Girlie
0:361:56Creative replacement with Flux KontextThe user demonstrates how to change a specific part of an image, such as replacing a building entrance with a glass door, using the Flux Kontext model in ComfyUI while maintaining the original scene context.Urban Decoders
5:200:53Change season to winter in architectural scenesThe demonstration shows how to use a text prompt to transform a summer architectural shot into a winter scene with snow on rooftops and trees using Flux Kontext.Urban Decoders
6:471:52Selective inpainting with crop and stitchUsing the 'ComfyUI inpaint crop and stitch' nodes, the user demonstrates masking a specific window area to insert a glowing LED digital facade.Urban Decoders
12:540:41Editing text on shop frontsThe demonstration shows how to use Flux Kontext to change specific text on a shop sign while maintaining the original font style and position.Urban Decoders
15:063:12Transform 3D scenes with Qwen2-VL (Qwen-Edit)CurrentThe creator uses the Qwen-Edit workflow to perform complex scene modifications, such as changing weather to rain or snow, while preserving the original 3D geometry and text.Matt Hallett Visual
20:221:07Fix text and faces using Crop and StitchA demonstration of a custom 'crop image' node to isolate specific areas like signs or faces, regenerate them at native resolution, and stitch them back into the high-res image.Matt Hallett Visual
7:201:48Edit image details with Flux KontextThe user shows how to load a reference image and use a text prompt to change specific features, such as changing a cartoon bunny's eye color to red.pixaroma
19:191:08Modify hairstyles and facial featuresThe creator shows how to change a subject's hair color and style (e.g., to a blunt bob with bangs) while attempting to maintain facial consistency using Flux Kontext.pixaroma
21:030:26Edit text within an imageA demonstration of Flux Kontext's ability to change text on a 3D render from 'Welcome' to 'Pixaroma' while maintaining the original font style and perspective.pixaroma