Gemini Omni 1.2 Flash AI Video Generator

Vertical 9:16, one uninterrupted 15-second take — no cuts, no on-screen text, no readable logos. 0:00–0:04 HOOK: Tight vertical framing on hands pulling the tab on a slim matte black cold-brew can; cold vapor wisps; golden-hour sidelight; shallow depth of field. 0:04–0:11: Smooth handheld ease-out to chest-up, eye level: Latino man, late 20s, faded navy henley, one gold earring, rooftop railing behind him — soft city skyline bokeh at sunset. He cracks the can (sharp carbonation hiss), takes a visible sip, rolls shoulders like he just clocked out, turns to camera with easy confidence. 0:11–0:15: Direct eye contact. He says in a dry, warm baritone: "Strong, cold, not sugary — that's the whole brief." Native audio: dialogue forward and dry; loud can-tab pop and fizz on open; swallow; distant traffic and wind — no music bed, no voiceover reverb. Look: photoreal skin pores and stable face through the turn; cinematic phone-capture energy with subtle handheld; teal-orange sunset grade; premium DTC beverage ad — polished but not broadcast-stiff.

Using the provided reference images for the café interior, the iced latte product, and the barista portrait style: Vertical 9:16, one continuous 8-second take in that same cozy morning café. Barista (match the friendly talent reference) walks into frame and sets the iced latte on the counter toward camera, natural smile, warm window bokeh behind her. Lighting and wood tones must match the interior reference; glass and drink must match the product reference — no readable logos on cup, apron, or signage. Native audio: soft café murmur, glass clink on counter, faint steam hiss, no music, no on-screen text. Single continuous social reel — not a slideshow or hard cuts between refs.