Qwen Image 2.1 Deep Tutorial: 20+ ComfyUI Workflows From Generation to Multi-Image Editing
pixaromadesign · reddit · 2026-09-30
Pixaroma released episode 36 of its ComfyUI tutorial series, demonstrating Qwen Image 2.1 with 20+ workflows, all included for free.
Text-to-image: different sizes, low-VRAM workflows, prompt enhancement, image-to-prompt, transparent PNG, Turbo LoRA.
Image editing (the model's strength): editing existing images, character sheets, inpainting, background removal, outpainting, style changes, plus the author's Sketch Pixaroma node for visually marking areas to edit.
Multi-image editing: virtual try-on, adding logos to products, placing characters in scenes, copying poses, combining up to six reference images.
The author shows both good and bad results to map the model's real limitations.
More from Multimodal
- Clueso MCP tops Product Hunt's Product of the Week for AI video editing — Scobleizer · 2026-09-30
- Flux 3 split-screen renders impress again, same prompt tested on Kling 4.0 Flash — umesh_ai · 2026-09-30
- YC launches Video Arena, a free arena where AI models compete to code your video prompts — ycombinator · 2026-09-30
- Mirage launches free Tesseract: agents design and animate editable visuals without Figma — chrisfirst · 2026-09-30
- Qwen image model's style comprehension falls short vs Krea 2 Turbo — Dear-Spend-2865 · 2026-09-30
- Automating a daily AI news video with Muse, Pi and GLM: a step-by-step workflow — op7418 · 2026-09-30