What problem does it solve?
Produces an iPhone-photo-looking first frame that wins the 3–5 second stay-or-scroll decision by locking composition, realism, and visual “anti-polish” before animation starts.
Core Features & Use Cases
- Canonical first-frame generation: Creates the creator’s believable, phone-captured face image that acts as the visual design gate.
- Three-layer prompting for UGC realism: Combines a fixed photorealism pre-prompt, a color-reference JSON, and a precise scene description with strong negative constraints (e.g., no text/letters).
- Visual reference chaining for multi-frame formats: Enforces sequential generation where later frames must reference frame 1 to prevent face drift across settings.
- Validation-oriented iteration loop: Guides repeated attempts (typically 2–4) using quality checks like skin texture, identity preservation, practical lighting, and real-world clutter.
Quick Start
Run the first-frame generation for your campaign so the canonical frame1.png is created and ready to chain into /animate and /b-roll.
