Kling V4.0 Image-to-Video: direct a clip with first and last frames
Kling V4.0 Image-to-Video animates a picture into a clip of up to 30 seconds, steered by a last frame and keyframes. A storyboard workflow, prompts and tips.
Animating a photo is easy now. Getting it to end where you want is the hard part: the cat should land on the sofa, not wander off frame; the door should be open at the end, not still closed. Kling V4.0 Image-to-Video answers that directly. You give it the first frame, and if you like, the last frame and a row of frames in between, and it fills the motion between them, for anything from 3 to 30 seconds, with synchronized stereo audio. Think of it less as a photo animator and more as a storyboard you hand to an animator.
Three ways to use it, from simple to directed
- First frame only. Upload a picture (JPG or PNG) and describe the motion, or leave the prompt empty and let the model decide. Good for a living portrait, a moving landscape, a product that turns.
- First and last frame. Add an end picture and the model travels from one to the other. This is the mode for a reveal, a before and after, or a transformation.
- Keyframes. Under the advanced settings you can add middle frames, in order. First frame, last frame and keyframes together are at most ten pictures, so up to eight stops in between when you use both ends.
Step 1: make the frames belong together
The model is good at motion and bad at miracles. If the first frame is a daylight street and the last one a different street at night, it has to invent a teleport. Keep the same subject, lens and light across your frames and change only what should move. The easiest way is to generate the first frame, then edit it into the next ones with an image editing model, so everything comes from the same picture. Our guide on image-to-video motion prompts has more on what to describe and what to leave alone.
Step 2: write the motion, not the picture
The frames already show what things look like. The prompt should say how they move and how the camera behaves. Three examples:
The camera pushes in slowly. The candle flame flickers, the curtain moves in a light draft, and the woman by the window turns her head toward the camera at the end.Time-lapse: the empty workshop fills with half-built furniture as the light moves from morning to evening. Static camera.One continuous shot following the paper boat down the gutter, under the bridge, into the river. Gentle handheld camera, the sound of running water and distant traffic.Step 3: pick length and shape
- Duration: a whole number of seconds from 3 to 30, 5 by default. Give the model time in proportion to the number of frames: eight keyframes in five seconds is a slideshow.
- Resolution: 720p by default, 1080p when the clip is final.
- Aspect ratio: leave it empty and the clip follows the first frame. Set 16:9, 9:16, 1:1 or 21:9 only if the destination needs it, and then crop your frames to that shape first so nothing important is cut off.
What it will not do
It does not start from nothing: a first frame is required. If you have only an idea, write it for Kling V4.0 Text-to-Video, or generate the first frame as an image. And it does not take character sheets or motion clips as references; for that, Kling V4.0 Reference-to-Video accepts images, videos and audio, explained in its guide. For a gentler start on animating pictures, read turn a photo into a video; every other image-to-video model is on the models page.
Questions
Do I need a prompt for Kling V4.0 Image-to-Video?
No. The first frame is the only required input. A prompt helps you choose the motion and the camera.
How many keyframes can I add?
First frame, last frame and middle keyframes together are at most ten pictures, so up to eight in between when you use both ends.
How long can the clip be?
Any whole number of seconds from 3 to 30, with 5 as the default, at 720p or 1080p.
