Kling O1 Edit — video-to-video editing from a prompt
Kling O1 is Kuaishou's unified multimodal video model. Its edit mode reads the motion structure of footage you already have and applies a written instruction across the whole clip — swapping a character, changing the environment, restyling the shot — while the camera movement and timing stay exactly as filmed. No mask to paint, no keyframes to set.
Upload a 3–10 second clip (MP4 or MOV, 720p–2160p, up to 200MB), describe the change, and generate. Sign up is free and includes starter credits. To animate a still photo instead of changing existing footage, use the AI Video Generator.
What Kling O1 Edit can change
Swap the character
Replace who is in the shot while their blocking, gait and screen position carry over. Attach reference images to pin the new identity rather than describing it in words.
Change the environment
Move the same performance to a different place. Because the camera path is preserved, a handheld push-in stays a handheld push-in in the new location.
Restyle the whole shot
Anime, claymation, film stock, night-for-day. Style is applied across the full duration, so it does not flicker frame to frame the way per-frame filters do.
Change wardrobe and props
Alter clothing, colours and objects without rotoscoping. Useful for variant versions of an ad from one filmed take.
How It Works
1. Upload Your Video
Upload a short video clip (3–10 seconds, MP4 or MOV). The video should be between 720p and 2160p resolution, max 200MB.
2. Describe Your Edit
Write a prompt describing what you want to change. For example: "Replace the background with a futuristic cityscape at sunset" or "Transform the character into a cartoon style."
3. Generate & Download
Hit Generate and Kling O1 will produce the edited video while maintaining the original motion and camera angles. Download when ready!
Reference images and tagged elements
A prompt alone leaves the model guessing what "a woman in a red coat" means. Attach up to four references in total to remove the guesswork — either plain reference images, or tagged elements.
An element pairs one frontal image of a person or object with up to three additional angles. That extra coverage is what holds an identity steady as the subject turns through the shot, which a single flat reference cannot do. Elements and plain images share the same budget of four.
Writing an edit prompt that lands
Name only what changes
The model keeps everything you do not mention. "Replace the office with a rain-soaked Tokyo street at night" works better than a full re-description of the shot, which invites it to alter things you wanted left alone.
Describe the target, not the operation
Write what the frame should contain, not the editing verb. "A grey wolf walking" beats "use AI to transform the dog into a wolf".
Keep lighting consistent with the new scene
If you move a daylight shot indoors, say so — "lit by warm practicals" — or the original lighting fights the new environment.
One change per pass on hard shots
Character swap and environment swap in a single instruction can trade off against each other. On a difficult clip, run them as two passes and keep the better intermediate.
FAQ
Is Kling O1 Edit free?
Sign up is free and comes with starter credits, so a first short edit costs nothing out of pocket. After that, edits bill per second of output — see the cost below.
How much does it cost?
Kling O1 Edit costs 5 credits per second of video. A 5-second edit costs 25 credits (~$2.50), and a 10-second edit costs 50 credits (~$5.00).
What video formats are supported?
MP4 and MOV files, 3–10 seconds in length, 720p to 2160p resolution, up to 200MB. Cut longer footage into single shots first — one edit per shot also gives the model a more consistent scene to work with.
Can I control who appears in the edited video?
Yes. Attach up to four reference images or tagged elements. An element pairs a frontal image with up to three extra angles, which is what keeps an identity consistent as the subject moves.
Can I keep the original audio?
Yes! Toggle the "Keep Audio" option to preserve the original audio from your source video. Turn it off when the edit changes the scene enough that the source sound no longer matches.
What is the difference between Kling O1 Edit and image-to-video?
Edit mode starts from a video and rewrites its content while keeping the motion. Image-to-video starts from a still and invents motion — for that, use the AI Video Generator.