Hand-held unboxing ad in 9:16: 8 seconds of Omni plus captions, $1.20

A hand-held unboxing ad in 9:16 costs $1.20 on Sume: an 8-second Gemini Omni Flash 1.1 clip at 720p is $1.00, and burned-in captions are $0.20. Ten variants with different openers cost $12.00.

This is the faceless version of a UGC ad. Hands, product, box and room sound, with no talking person. If you need someone speaking to camera, see the avatar route below.

What the prompt should say

Omni takes a text prompt plus image references. Give it the product photo as the first frame, then describe the camera like a phone: handheld, slight shake, kitchen table, daylight, hands opening the box. Omni makes native audio, so ask for paper and box sounds rather than music.

• Keep it 3 to 10 seconds. Eight seconds fits one unboxing beat.
• Use 9:16. Omni supports 16:9 and 9:16 only.
• Vary one thing per variant: the opening shot, the surface or the lighting.

Cost per ad

Resolution is the main lever. Captions are a flat per-job price, so they do not scale with the clip.

What to check frame by frame

Text on packaging is where generated clips fail most often: letters shift between frames or turn into near-words. Pause the clip on the box and the product label. If the brand name is wrong in any frame, the clip is not usable as an ad, however good the motion.

Hands are the second check. Count fingers on the frames where hands cross the box edge. A bad frame is a regenerate, and at 720p an 8-second retry is $1.00.

When you need a talking face

A person speaking to the camera is a different job. On Sume that is avatar video: a ready avatar plus a script, 4 to 60 seconds. The Standard tier is $0.184 a second, so 8 seconds is $1.47, before captions.

Be plain with viewers. A generated hand is not a customer, so do not present the clip as a real buyer's unboxing. Check the ad policy of the network you run on.