GPT Image 2 prompt recipe: motion-trail hand-drawn annotations on photos
songguoxiansen · x · 2026-08-27
A detailed GPT Image 2 editing prompt: overlay hand-drawn annotations on a photo while strictly preserving the subject, pose, clothing, props and original lighting. The "motion-trail composition" lets acid-green marker annotations grow naturally along the subject's movement, a pet's gaze or object tilt. The prompt specifies the main annotation cluster near the action focal point, three consecutive trail segments, a secondary beat-point cluster, 25%-30% quiet space, and explicit bans on adding nonexistent objects, neon/cyber UI, or occluding faces, hands and logos.
More from Multimodal
- H3 Fun ControlNet Brings Control-Video Guidance for MiniMax-H3 to ComfyUI — wyzborrero · 2026-08-27
- Sopro V2: SOTA-level 120M param voice cloning TTS — SammyDaBeast · 2026-08-27
- Achieving World Consistency in Video using 360 Panoramas and Minimax H3 — HighGaiN · 2026-08-27
- Minimax H3 Config Guide for 16GB VRAM: Balancing Speed and Quality — tj-tj-tj-tj · 2026-08-27
- video-use + coding agents for video editing: "one of the best AI tools I've ever used" — jacob_posel · 2026-08-27
- Golden age of artificial world building: imagegen capabilities are truly amazing — MarvinTBaumann · 2026-08-27