AI UGC Video Generator: One Outfit Photo to a 25s Pick-Up OOTD Reel

What Ribbi did

Ribbi feeds the reference into Nano-Banana-Pro's character-consistency mode and delivers a 1536x2752 vertical outdoor keyframe: on a European cobblestone street, the navy polo-collar knit, light-blue washed wide-leg jeans, white western belt, gold chain, and aviator sunglasses are reproduced piece by piece — a 90s retro-collegiate, old-money street look, with the model's face and hairstyle fully aligned to the reference.

What Ribbi did

Ribbi first runs the keyframe through the virtual-asset channel to obtain an asset ID that clears Seedance 2.0's strict real-face review, then drives image-to-video: 15 vertical seconds cycling full-body, low-angle, waist-level tracking, and walking push-in shots as the model works the complete outfit down a European street in harsh midday light and film grain. Ribbi then stitches Acts 1 and 2 into a full-length rhythm preview so the user can check the two-act structure end to end.

What Ribbi did

Ribbi rewrites the timeline prompt as a single continuous uncut shot: a fixed camera pinned about 20 cm above the floor, framing only below mid-calf, while the model's hand lifts the scattered knit, jeans, belt, and duffle bag piece by piece, then takes three steps out of frame leaving only shadows on the floor — 10 seconds in one take, fabric rustle and heavy footsteps carried entirely by SFX. This version passes on the first look and goes straight into the final cut.

What Ribbi did

Ribbi generates an old-school boom bap instrumental with MiniMax — heavy kick and snare, vinyl crackle, a dark fashion-editorial mood — and plays it for the user first; after a one-word approval it mixes BGM at 0.5 against ambient SFX at 0.6 and recomposes the 1080x1920 vertical 25-second final: the 10-second pick-up take hard-cuts into the outdoor multi-shot OOTD, clean and tight, accepted on the spot.

What Ribbi did

Ribbi proactively writes two field-tested lessons back into the skill: Act 1 locked to a 10-second single continuous take, and Act 2's keyframe required to run character-consistency with the reference image fed in, with the default spec updated to 25 seconds total (10s + 15s). The next outfit photo can run straight through without hitting either pitfall — one session distilled into an infinitely reusable OOTD template.