oleshevakatya/ComfyUI Qwen Edit Workflow for Head Swap no LoRAs needed!
Finally sharing my stable, production-ready head swap workflow for ComfyUI. No custom LoRA training needed. What it does: Swaps heads from source → target image while preserving pose, lighting, and scene context Runs on Qwen Edit + Lightning LoRA (4 steps) – super fast Automatic face masking, cropping, and intelligent blending Smart system prompts generated from both images via vision models ControlNet + depth/structure for alignment control Final upscale with SeedVR2 for clean, polished output Key features: ✅ Fast (4-step Lightning LoRA) ✅ Automatic mask generation refinement ✅ Pose + express
| Author / organization | oleshevakatya |
|---|---|
| Downloads | 3.9K (as reported by Civitai) |
| Likes | 121 |
| Task | text-to-image |
| Created | 27 November 2025 |
| Last updated at source | — |
Popularity, last 3 days
3,867 now (+8 since 2026-10-15). Counts as reported by the source, snapshotted daily by Silytics.
More from oleshevakatya
Silytics shows public metadata from Civitai and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.