How the KEPLER headphone review was made with no reviewer and no product: two reference images, a hero frame with the product already on the desk, one timed prompt, one video model, all through one connector. The exact prompt is below, ready to paste.
Open Nano Banana Pro (gemini-3-pro-image) and make two stills. One: the product on seamless white in a three-quarter view so every surface and the brand mark are visible (for a client, use their real packshot). Two: a phone selfie of the creator in the room where the review happens, written like a casting note: hair, beard, earring, sweatshirt, the desk behind him. Add the realism layer every time: visible pores, flyaway hairs, no smoothing, subtle sensor noise, muted natural grade. Fictional brands only when you are practising.
02
02Step 2
Put the product in frame one
Make one more image with both references attached and say what each is for: reference 1 is the product, reproduce it exactly; reference 2 is the person and the room, identity only. Describe the opening frame: him at the desk, the headphones in front of him, one hand already on them. This is the start frame of the video. Because the product is already in the frame, the model never has to invent it. Fix the product here, not in the video.
03
03Step 3
Write the prompt as a shot list
Copy the prompt below and change only the product, the person and the four lines. Keep the shape: an identity lock at the top, four timed beats (hook, first look, wear, payoff), each with one camera position, one action and one line in quotes, then a style line and a negative line. The honest-take opener ("I've had these for a week") is the review format that gets watched. Stay under 300 words.
04
04Step 4
Run it through one connector
In Claude Code, the ElevenLabs connector gives you the image model and the video model behind one login (add it with claude mcp add --transport http elevenlabs https://api.elevenlabs.io/v1/mcp). Wire the hero frame into Kling 3 Pro as the start frame, 15 seconds, 16:9, native audio on, and ask for the estimate first. The 15-second clip is about $3 in credits; every still is about a quarter. Expect the render to take longer than the estimate says.
05
05Step 5
QC like an editor, then post the breakdown
Watch the clip three times: once for the face across the cuts, once for the product (same cups, same leather, same mark), once for the hands. If one shot drifts, regenerate the still for that shot and re-run. Then post the breakdown, not just the ad: the clip on top, the two references under it, the prompt scrolling while it plays. Same prompt shape made a serum ad, a sneaker unboxing and this review in one day: change the two references and the four lines.
The exact video prompt (copy it, change the product, the person and the four lines)
Create a realistic 15-second UGC tech review video using the provided man as the consistent main character and the provided KEPLER over-ear headphones as the exact product reference. Keep his face, short dark hair, short beard, silver stud earring and black crewneck identical throughout. The headphones keep their exact shape, matte sand-beige cups and headband, dark brown leather cushions, brass hinges and the small KEPLER mark on the right cup. Do not redesign, recolor or alter the product.
0:00-0:03 HOOK. Stationary iPhone propped on the desk, facing him. He sits at his desk in morning light with the headphones resting on the desk in front of him, one hand already on them. He leans toward the lens. Dialogue: "Okay, I've had these on for a week. Here's the honest take."
0:03-0:06 FIRST LOOK. Hard cut, same camera. He lifts the headphones and turns them slowly so the right cup with the KEPLER mark and the leather cushion face the lens, thumb pressing the cushion. Dialogue: "First thing: this padding. Actual leather. It doesn't get hot."
0:06-0:11 WEAR. Hard cut to a slightly tighter stationary angle. He puts the headphones on, adjusts the band with both hands, taps the right cup once and pauses, eyes drifting off as he listens. Dialogue: "And then the sound just... closes the room."
0:11-0:15 PAYOFF. Hard cut. He slides one cup off his ear, looks straight at the lens with a small grin and nods. Dialogue: "Yeah. These are staying on the desk."
Visual style: iPhone camera quality, natural window daylight with a warm desk lamp, slightly imperfect exposure, real room ambience, visible skin texture, no filters, no studio lighting, not a commercial. Audio: his natural voice, faint room tone, a soft click when he taps the cup. No morphing hands, no extra people, no on-screen text.
Want this running for you?
Build it step by step inside AI Creative OS, with me in the thread.
