Built for cinematic video generation with native audio, multi-shot scenes, and strong character consistency, Kling 3.0 gives you serious control over how your video looks, moves, and sounds.
How to generate a video with Kling 3.0
Add a Create Video Flow to the Canvas
Click Create Video on the toolbar on the left of the Canvas. Click the model name to open the model menu and select Kling 3.0 Standard Video.
Add prompts and images
Subject prompt: Describe the full scene.
Start frame (optional): Upload or drag in an image to anchor the opening shot. The model preserves the look, layout, and details of your image while adding motion.
Start and end frames (optional): Define both the first and last frame for precise control over how the video begins and ends. Great for transitions, reveals, and creating looping videos. How to Create Looping Videos in Kaiber Superstudio
Adjust settings
In the Advanced Features section of the Create Video Flow, set:
Duration: 3 to 15 seconds
Aspect ratio: 1:1, 9:16, or 16:9
Audio: Toggle on for native dialogue, sound effects, and ambient audio. Toggle off if you're adding your own audio later.
Generate
Click Generate. From there you can upscale, download, extend, restyle, or use your video in a Video Editor project.
Prompting for Kling 3.0
In your prompt describe the scene, camera movement, action, lighting, and sound as a flowing instruction. For example: A rainy city rooftop at night, a woman in a dark coat turns toward the camera, wind catches her hair, slow dolly push-in, distant traffic sounds, moody blue lighting.
Prompting for Multiple Shots
Kling 3.0 supports multi-shot generation. Label each shot clearly and describe the framing, action, and camera for each one.
Shot 1: Medium close-up of a man sitting at a desk in a dimly lit office, typing slowly on an old black typewriter, warm lamplight, static camera.
Shot 2: Wide shot of the city skyline through the window, rain streaking the glass, ambient traffic sounds.
Keep Characters Consistent
Using a start frame image of your character will give control over the character's appearance and help maintain consistency through the video.
To maintain the same character across shots, assign each character a clear character label and description at the start of the prompt. Be specific about key features like clothing, hair, and build. Use the label to refer to the character throughout the prompt.
Character A is an caucasian man age 60 with the disheveled hair of an eccentric professor, wearing a tweed jacket
Shot 1: Medium close-up of Character A sitting at a desk in a dimly lit office, typing slowly on an old black typewriter, warm lamplight, static camera.
Shot 2: Over-the-shoulder shot of Character A as he pauses and looks toward the window, camera slowly pulls back.
Generating videos with audio and dialogue
Kling 3.0 can generate dialogue, ambient sound, and sound effects synced to the action.
To create dialogue, tag each speaker.
[Character A: The detective, calm low voice]: "I wouldn't do that if I were you." [Character B: The bartender, nervous whisper]: "You didn't hear it from me."
For ambient sound, describe it as part of the scene: Rain hitting the window, distant traffic sounds.
Kling 3.0 supports multilingual dialogue including English, Chinese, Japanese, Korean, and Spanish with accent options.