What problem does it solve? Text-to-video generation from scratch often produces unrealistic results with identity drift and artificial-looking motion. This Skill solves that by using a still-then-animate pipeline: generate or reuse a realistic anchor image first, then animate it with focused motion prompts inside Palmier Pro. ## Core Features & Use Cases - Still-then-animate pipeline: Locks subject identity, lighting, and skin texture with an anchor still before adding motion, so the video model only solves movement and speech. - Model selection guidance: Picks between Grok Imagine Video, Kling V3/O3, Seedance 2, and Veo 3.1 based on dialogue needs, reference counts, resolution, and observed real-world failure modes. - Beat-based ad assembly: Splits a UGC ad into short jump-cut beats (hook, demo, CTA), locks voice consistency across clips with explicit pitch/tone/accent/pace descriptions or reference audio, and assembles clips on the timeline. - Use Case: A marketer animates a skincare product still into a talking-head hook clip on Kling V3 with a quoted line and handheld camera shake, then cuts it together with demo and CTA beats on the Palmier Pro timeline. ## Quick Start Ask the agent to animate an existing still from the media library into a short UGC-style talking clip with a specific spoken line and natural handheld camera motion.
