Start → end frame

You choose both ends.
The model shoots the middle.

Upload the frame the shot opens on and the frame it lands on, describe the move, and get one seamless generated shot that connects them — with your composition guaranteed at both ends, because you supplied it.

Video paused while off screen

One continuous move from your first frame to your last.

How it works

Two stills become one move

The most controllable shot in generative video, because you author its first and last frame.

1

Pick the two frames

The shot you want to open on and the shot you want to land on — product closed and product open, empty room and dressed set, logo apart and logo assembled.

2

Describe the journey

A sentence about how one becomes the other: the camera move, the speed, the mood. The model generates the motion that connects your two stills.

3

Edit it on the canvas

The result arrives as a scene node in the studio, so you can re-prompt it, trim it, or slot it between a presenter hook and a motion-graphics close.

What it's for

The cuts this replaces

Product reveals

Before and after, without the jump cut

Packaging closed to packaging open, ingredient to finished dish, sketch to shipped product — the transformation shot that normally needs a rig, a motion-control pass or an expensive 3D artist.

Video paused while off screen
Scene bridges

Connect two shots you already have

Take the last frame of one scene and the first frame of the next, and generate the camera move between them — a bespoke transition instead of a crossfade, built from your own footage's endpoints.

Video paused while off screen
Why it holds up

Control is the feature

Guaranteed endpoints

The shot must begin and end on your images. The model cannot drift the product or reframe the scene at either boundary.

A node, not an orphan

The result lands on the studio canvas as an editable scene, so it composes into a full ad with presenters, motion graphics and narration.

Priced like any scene

One scene-rate render in credits, shown before you commit, refunded automatically if the generation fails.

Questions

The things people ask first

You control both endpoints. A text-to-video prompt gives the model the whole shot to invent; here it must start exactly on your first image and land exactly on your last, so the product, framing and composition you chose are guaranteed at both ends.

Pick two frames and watch them connect

First-frame to last-frame generation, editable on the studio canvas. Free to start.

Start free