UGC workflow
One creator, one room, one activity, and a single unbroken take. Generate every reference first — her face, the room, and each product in the state it has to start in — then animate them in one generation.
The rule above all. One clean video is one simple location with one main activity, everything present from the first frame.
0 of 4 steps ready
01. Lock the creator
The most important asset. This locks her face and puts her in the brand's look. It is attached to every generation afterwards, so everything downstream inherits it.
The prompt
Still needs look, outfit, room
Hyperrealistic selfie-style photo. She wears . She holds the phone selfie-style in . Natural realistic glowing skin, hyperreal detail, vertical 9:16. No text.
The anti-slop checklist
The whole method on one page. Run it before you generate, not after you have paid for a render.
- Character locked to one reusable reference, in the brand's look, attached to every generation.
- Every product piece and the room generated in advance as its own reference.
- Product references shown in the right state — a bag's shoe compartment closed, not open.
- Camera choice made: selfie to open, then propped hands-free for the activity.
- State one continuous take, no cuts, time never skips.
- The full room described so no furniture duplicates or spawns.
- Everything she handles is laid out in the room from the first frame.
- Each item named precisely and told to stay that item, never morphing.
- Every object gets a complete action with a clear end.
- Dialogue is spoken to camera, not voiceover.
- No phone appears after it is propped down.
- A final locks block repeats every hard rule.
Ask ATLAS
It already knows these rules — the character lock, the room lock, one unbroken take, vertical 9:16. Ask it to write a field and it writes it ready to paste, and tells you what is missing first.