Stop wasting hours on manual face swaps. This workflow builds production-ready face models from a photo batch — automatically.<hr /><h2 id="what-this-workflow-does">✨ What This Workflow Does</h2>Feed it 20+ photos of any person. It automatically:<ul><li>Detects &amp; crops every face with sub-pixel precision</li><li>Upscales each crop to 512×512 using Bicubic interpolation</li><li>Extracts deep face embeddings via ReActor</li><li>Saves a ready-to-use <code>.safetensors</code> face model directly to your InsightFace folder</li></ul>One click. One model. No manual work.<hr /><h2 id="why-this-workflow-crushes-everything-else">🚀 Why This Workflow Crushes Everything Else</h2>Feature This Workflow Manual Method Setup time~5 minutes Hours Face extraction Auto-batch One by one Output format<code>.safetensors</code> (ReActor-ready)Varies Face alignment✅ Automatic❌ Manual Batch processing✅ Full folder❌ Image by image Node layout Clean, aligned grid Spaghetti chaos<hr /><h2 id="what's-included">📦 What's Included</h2><ul><li><code>FaceModel_Pro.json</code> — Drop-and-go ComfyUI workflow</li><li>Full EN Documentation (10-section professional guide)</li><li>README with Quick Start &amp; Tips</li><li>CHANGELOG with full version history</li><li>LICENSE — Clear personal &amp; commercial use terms</li></ul><hr /><h2 id="requirements">⚙️ Requirements</h2>Extensions (install via ComfyUI Manager):<ul><li><code>ComfyUI-ReActor</code></li><li><code>WAS Node Suite (was-ns)</code></li></ul>System:<ul><li>Python 3.10+</li><li>ComfyUI (latest)</li><li>CUDA 11.8+ recommended</li><li>8 GB RAM (16 GB recommended)</li><li>~500 MB disk for models</li></ul><hr /><h2 id="quick-start-3-steps">🎯 Quick Start — 3 Steps</h2>Step 1 — Install the two extensions above via ComfyUI ManagerStep 2 — Drag &amp; drop <code>FaceModel_Pro.json</code> onto your ComfyUI canvasStep 3 — Set your image folder path → set your model name → hit Queue Prompt✅ Your face model saves to: <code>ComfyUI/models/insightface/</code><hr /><h2 id="pro-tips-for-best-results">💡 Pro Tips for Best Results</h2><ul><li>Use 20+ images minimum — more photos = stronger, more consistent model</li><li>Good lighting is everything — avoid heavy shadows or backlit shots</li><li>One face per frame — multi-face images confuse the detector</li><li>Increase <code>crop_padding_factor</code> to 0.35–0.5 if you want neck/shoulders included in the crop</li><li>Consistent expressions across your dataset improve embedding quality</li></ul><hr /><h2 id="workflow-defaults">🔧 Workflow Defaults</h2><pre><code>crop_padding_factor : 0.25
upscale_method : bicubic
output_resolution : 512 × 512
save_mode : ON
select_face_index : 0
batch_mode : incremental_image</code></pre>All values are fully adjustable inside ComfyUI.<hr /><h2 id="roadmap-coming-in-v1.1.0">🗺️ Roadmap — Coming in v1.1.0</h2><ul><li>Live Preview Image node before saving</li><li>Multi-face export per batch run</li><li>LANCZOS upscale option</li><li>Face quality filter — auto-skip blurry or low-res detections</li><li>Arabic documentation</li></ul><hr /><h2 id="faq">❓ FAQ</h2>Q: Does this work with anime faces? A: Detection uses <code>lbpcascade_animeface.xml</code> — yes, anime faces are supported!Q: Can I use the generated model commercially? A: Check the included <code>LICENSE.txt</code> for full usage rights.Q: What if no face is detected? A: Make sure images are well-lit and front-facing. Increase <code>crop_padding_factor</code> slightly and retry.Q: How many photos do I actually need? A: 20 is the minimum. 40–60 clean, varied photos will give you a noticeably stronger model.<hr />

🧠 ComfyUI Face Model Pipeline — Powered by ReActor | Build Perfect Face Models in 3 Clicks

📝 Versionv1.0.0 — March 2026 (Initial Release)If this workflow saved you time, hit that ❤️ and drop a ⭐ — it genuinely helps more people find it. Questions? Leave a comment below — I check daily.

Files with SHA256: c27a5c3477c2e893effe28d2eaf9f79240e7086a902d5fb9034b2ab54ffdeab1