Bunsy · iOS & Android
First and last frame, one shot out
Set the start, set the end, get the move.
Or open the download link
Models, chained
First and last frame generation gives a model two visual references: where the shot starts and where it should end. Use compatible frames to guide the transition, then review how the model connects them.
Direct motion with a first and last frame
A prompt asking for an orbit returns whatever that model thinks an orbit is. Two frames put the start and the end beyond argument, and leave the interpolation as the only open question.
Ferrari Testarossa 1980
Ferrari Testarossa driving along a Sicilian coastal cliff road at golden hour, cinematic shot.
Why it holds up
Continuity across the cut
Use the closing frame of one shot as the opening frame of the next and the edit has something to hold onto. It is what stops a sequence reading as unrelated clips of a similar subject.
Either frame can be generated
The endpoints do not have to be photographs. A still you made with the image models, a frame pulled out of an earlier clip, a photo from your camera roll — anything in the library can anchor an end.
Made in Bunsy
What it produces
Clips made in Bunsy, exactly as they came out of the app.
A man on a Manhattan street
An anime kid running through tall grass
A mug of coffee at sunrise
Running over a fallen tree in a forest
A woman reaching towards a park fountain
A princess at a mirror, generated in Bunsy
A robot watching over an editor at work
Questions
What makes a first and last frame pair work?
They need to read as the same scene — same subject, same light, a plausible distance apart. Two unrelated images get you an invented transition rather than a shot, because there is no move that honestly connects them.
Which models take a first and last frame?
Several in the catalogue accept two inputs rather than one, and the picker shows which as you choose. A frame pair that comes out wrong on one model is usually worth retrying on another before you change the frames.
Is this better than writing the camera move into the prompt?
For the move itself, yes — you are handing over an answer instead of a description. The prompt still governs what happens during the move: the light shifting, the subject acting, the speed of it.
Your first clip is
a minute away.
Free to download. Nothing to set up.