SocialSight supports GPT Image 2 and Seedance skincare ad prompts with locked references

• AIwithSynthia's commercial prompt shows a skincare ad workflow that locks the model's face, wardrobe, product packshot, skin texture, color palette, aspect ratio, and music direction before any video render starts.
• In AIwithSynthia's shot list, that master prompt turns into an eight-shot production plan with lens choices, clip lengths, motion notes, and sound design cues, from alarm-clock closeups to a final pedestal product shot.
• AIwithSynthia's fashion test uses the same pattern on a still image, preserving a portrait subject's exact likeness while changing scene, outfit color, and editorial framing.
• SocialSight's homepage frames the product around reusable assets and studio presets, while the platform link in the shot-list post describes saved characters, products, and styles as a way to keep outputs on-brand across image and video.

You can browse SocialSight's homepage, skim its prompt guide, and the linked platform overview makes the core pitch unusually explicit: save a character, product, or style once, then reuse it across generations. The interesting bit in the main skincare post and its follow-up thread is how much of a real ad brief now fits inside that system, down to peach fuzz, linen rustle, and the exact color blocking on a La Roche-Posay tube.

The prompt in AIwithSynthia's skincare commercial prompt reads less like a one-off text prompt and more like a brand bible. It fixes five references at once: subject identity, wardrobe continuity, skin realism, product packaging, and soundtrack tone.

That structure matters because the prompt is not only telling the model what to make. It is also telling it what must not drift. The negative constraints are as detailed as the positive ones:

• same female model across all shots
• loose cream shirt and high-neck inner layer, no exposed shoulders or chest
• visible pores, peach fuzz, and under-eye texture, no glossy or waxy skin
• exact tube colors, typography, and cap shape for the sunscreen pack
• soft premium music, never loud or trailer-like

AIwithSynthia's fashion editorial prompt applies the same logic to a still image. The subject's facial features and hairstyle stay fixed, while the environment, wardrobe color, and gallery-scale mural become the variables.

The follow-up in AIwithSynthia's eight-shot breakdown turns that reference lock into a literal preproduction document. Each shot gets a focal length, camera move, duration, environment, and audio bed.

The sequence breaks down like this:

1. 85mm sunrise bedroom push-in, alarm tap, linen rustle
2. 50mm bathroom mirror setup, product hero in foreground
3. 100mm macro pump shot, cream texture and packaging colors
4. 85mm side-profile application pass, sunscreen disappearing into skin
5. 100mm macro skin texture hold, pore detail and matte-hydrated finish
6. 35mm wardrobe and outdoor transition, shirt button and city walk
7. 85mm outdoor portrait, protected skin in direct sunlight
8. 85mm centered pedestal packshot with final brand line

The video attached to the main post is the payoff. The notable part is not just that it moves, it is that the motion clip inherits the same continuity rules the text already set up.

The official SocialSight homepage says the platform is built to "create AI images and videos that stay consistent," then describes reusable assets for characters, products, and styles, plus studio presets that turn a workflow into a repeatable app. The summary attached to SocialSight makes the same point in plainer language: save identities once, reuse them across images, videos, and templates.

That framing fits the evidence tweets closely. The skincare workflow treats the model, the sunscreen tube, and the visual style as stable assets. The shot list treats the ad itself like a preset, a sequence that could be rerun with the same references and a different product claim or scene.

SocialSight's prompt guide is generic, but it reinforces the operating model here. Detailed prompts are not filler copy. They are the control layer for a workflow that wants brand consistency to behave like a saved production asset.