A full GPT Image 2.5 workflow for making UGC ads
The author calls UGC ads the easiest thing AI can make right now and shares a full GPT Image 2.5 plus Seedance 2.5 workflow with every prompt.
save this one.. UGC ads are the easiest thing AI can make right now and almost nobody is doing it properly full workflow, every prompt, GPT Image 2.5 + Seedance 2.5 👇
2/ the character sheet one photo in, six angles out. GPT Image 2.5 holds the face across all of them now, that's the whole reason this works. Prompt: Multi-panel phone-camera character reference sheet of the same real woman across six panels: front-facing, three-quarter left, full left profile, full right profile, three-quarter right, and a slight chin-up front view, all shot against the same plain neutral beige wall in ordinary flat indoor lighting. Real everyday face, not a model: fair skin with visible uneven tone, natural pink flush across the cheeks and the bridge of the nose, scattered faint freckles, small blemishes and a little redness around the nostrils and chin, visible pores across the nose and cheeks, an oily sheen on the forehead and nose bridge, fine peach fuzz along the jaw and cheek edge, faint under-eye shadow and mild puffiness, slightly asymmetric features with one eye a touch different from the other. Pale green-grey eyes with real iris fiber detail and wet catch-lights, natural full light-brown brows with visible individual hairs and small gaps, straight nose, naturally full pink lips with fine vertical texture and slight dryness. Light brown hair with soft golden tones, unwashed and slightly flat at the roots, pulled back loosely and messily with several loose frizzy pieces escaping around the temples, ears and neck. She wears a plain grey marl cotton t-shirt, soft and slightly worn, with a loose stretched neckline. Identical lighting, identical skin, identical hair and identical shirt in every panel. Flat ordinary indoor light from a ceiling fixture, slightly warm, small soft shadows under the eyes and nose, no window glow, no rim light, no styled lighting. Shot on a smartphone camera: mild digital noise, slight softness from noise reduction, visible compression texture, natural handheld framing that is not perfectly level or centered. No beauty filter, no skin smoothing, no airbrushing, no retouching, no makeup, no bokeh, no cinematic grade, no studio backdrop, no model bone structure, no perfect symmetry, no flattering angles.
