Show HN: Wan-Animate – Unified Character Animation and Replacement (www.wananimate.net)

Wan-Animate is a new unified framework for animating and replacing characters in video by transferring a performer’s motions, expressions, and camera dynamics onto any static character image. Using what the authors call “holistic replication,” the system supports arbitrary characters (from anime avatars to photoreal humans), reproduces facial expressions, body motion and camera perspective, and integrates relighting so replaced characters preserve scene lighting, shadows and color tone. Key technical components noted include motion transfer and a built-in Relighting LoRA (low‑rank adaptation) to blend characters naturally into environments, plus an end-to-end workflow for upload, reference selection, animation/replacement and preview. The project is also shipping model weights and code, enabling community extension. For the AI/ML community this matters because it bundles animation and replacement into a single, usable pipeline with open weights—accelerating research on character motion synthesis, relighting, and cross-domain generalization while lowering the barrier for creators. Practical applications span content creation, games, education and VFX, but the release also raises safety and provenance concerns (deepfakes, copyright and consent), highlighting the need for watermarking, detection tools and responsible-use policies alongside continued model audits and technical safeguards.