Image references
Choose the supported inputs and model-aware controls in one focused AllVideoAI workflow.
Kuaishou · Video model
Create with Kling O3Kling O3
A reference-aware Kling model for guided subjects, scenes, styles, and creative transformations.
Model overview
Meet Kling O3
Kling O3 is a unified multimodal video family built for text, image, reference, and editing workflows. Its reference modes can extract characters, props, or scene identity from several views and carry them into a new composition.
The same family also supports prompt-led video edits such as replacing objects, changing a background, restyling a scene, or adjusting weather and light while aiming to preserve temporal consistency.
Video references
Choose the supported inputs and model-aware controls in one focused AllVideoAI workflow.
Video transformation
Choose the supported inputs and model-aware controls in one focused AllVideoAI workflow.
Prompting tip
Name each uploaded reference in the prompt and explain what must be preserved versus what may change. For edits, describe one clear transformation at a time.
AllVideoAI Studio
Compare models and create in one workspace
Supported creation modes and parameters adapt to the selected model, with credit pricing confirmed before generation.