Clear faces or features
Choose authorized photos with good light and no heavy filters. Keep each role visually distinct.
AI TRENDS / REVEALS & REACTIONS
Cast two people in a fictional kitchen gag. The bright soft prop recoils harmlessly toward the prankster’s own hands, followed by a delayed glance and an awkward recovery.
Generated videos are stored for 24 hours, then deleted. Download files you want to keep.

Rubber Band Prank AI · about 9s · 768 × 1024 original · Sound on play. Compact preview; native 768p sources available below.Recorded source demo · MiniMax H3
Open the video editor with the prepared prompt. Review credits before generating; guest checkout is available.
Original EveryGen AI demo generated with MiniMax H3 at 768 × 1024 with generated audio. The green prop snaps back to the prankster’s hands. The companion continues at the counter while the prankster holds the awkward reaction. The Template uses the reviewed silent motion with generated sound on a new take.
Template photo roles: Prankster, Companion. Normal supports editable prompts and compatible native inputs.
Use my two photos for the left-side prankster and right-side companion. Keep the soft green prop with the prankster and end with a delayed sideways look.
Normal generation lets you edit the starting prompt and choose a model. Template follows this example using MiniMax H3 at 9s / 3:4 / 768p with audio. Review the quote before generating. Original prompts do not guarantee identical results.
THE SCENE / YOUR CREATIVE DIRECTION
Cast two people in a fictional kitchen gag. The bright soft prop recoils harmlessly toward the prankster’s own hands, followed by a delayed glance and an awkward recovery.
The prankster and the bright prop are already visible on the left.
The prop retracts toward the prankster’s own hands.
The companion’s glance lands after the motion; hold the awkward beat.
This separate video guides the action, camera and timing. Your subject photos change the featured character; this theme preserves its existing setting.
Prepare the identities and framing before opening a new take in EveryGen AI.
Choose authorized photos with good light and no heavy filters. Keep each role visually distinct.
Understand what the source brief controls before you edit it in EveryGen AI.
Cast the prankster and companion separately and preview their positions before generation.
The prop stays with the prankster. The second actor supplies a look or small reaction.
An open two-shot helps you read the setup, recoil and payoff without losing either actor.
Start in Inspiration to shape your idea, or choose Normal to edit the prompt and model. Template follows the example. Review the quote before generating.
Place the prankster in the first slot and the companion in the second. Use distinct, authorized photos.
Check the left and right roles and review the live cost for the template.
Play back the whole clip and look for continuous hands, stable identities and a reaction after the prop moves.
Use the finished video to choose one useful change for your next EveryGen AI take.
Review the full clip, especially the middle transition and ending. Simplify the action in Normal or supply a cleaner photo for a difficult role.
Prompts, reference inputs, source models and your new video.
This template requires both named photos. In Normal, you can reference one subject and ask the selected model to invent a fictional companion.
In Template, the photos guide the two appearances. The scene follows a reviewed silent motion reference.
It is designed as staged comedy with a soft visual prop and a startled reaction. The prop should stay with the prankster and avoid contact with the face or the other actor.
Check the live quote before starting. It reflects the selected model, output time, included reference seconds and any chargeable inputs. Normal settings and Template may have different quotes.
Use Normal to edit the starter prompt and compatible native settings. Template follows the reviewed example's movement and assigns your photos to named roles.
Track the task in Creations and download the private video before its 24-hour availability ends.
Published EveryGen AI examples explain the inputs, settings and limits for different scenes.
Stage a readable one-take AI video around a single action, distinct photo roles and a clear ending.