Kuaishou
First and last frame transition specialist. An image-to-video model built around keyframes: give it a start frame, optionally an end frame and a runtime between 3 and 15 seconds, and it works out the motion in between. Audio is generated in the same pass when you ask for it.
This exact endpoint is not in the Studio catalog yet — the button opens our image-to-video workflow, where Kling 2.6 Pro and Kling 2.5 Turbo are available today.
Features
The start image is required and the end image is not. Supply both and the model interpolates a deliberate transition between them — a morph, a scene change, a before and after — rather than inventing an ending. Supply only the start frame and it animates forward from there.
The prompt is optional. When you write one it steers how the model travels between the two frames, covering pacing, camera behaviour and the look of the intermediate motion, while the keyframes stay in charge of where the shot starts and stops.
Multi-prompt takes an array of prompts and the shot type decides how they are cut: customize follows the shot breakdown you describe, intelligent lets the model place the cuts. One call produces a sequence instead of a single continuous take.
Audio generation is a single boolean and it is priced separately, so silent drafts stay cheaper while you iterate on framing and you pay the audio rate only on the take you keep.
Runtime is any length from 3 to 15 seconds rather than a fixed menu of clip lengths, so a three-second transition is billed as three seconds instead of being rounded up.
Examples
The character walks forward slowly, with the camera following from behind.
Try this promptStart frame to end frame: the closed product box opens and the contents rise into view.
Try this promptMorph the daytime street into the same street at night, holding the building line steady.
Try this promptThree shots: a wide establishing view, a slow push toward the subject, then a close-up on the hands.
Try this promptKling O3 Pro renders each example on demand rather than serving fixed sample clips, so prompts are shown instead. They are starting points for the transition, morph and multi-shot cases the model is built for.
Workflows
Cost is charged per second of output, and turning audio on moves the job to a higher per-second rate. Credits come from one shared balance across image, video and speech, and the exact estimate is shown in the Studio before you generate.
Built for
FAQ
Limits, audio, licensing and price — the things worth knowing before you spend a credit.
Open the Studio, load a start frame, and let the image-to-video workflow fill in the motion.
