Back to models
Kuaishou Technology (Kling AI)
Video modelHotKling Video 3.0 Omni
Kling Video 3.0 Omni is a model in the Kling 3.0 family that places greater emphasis on reference material and narrative control. Alongside starting from text or images, it suits projects with existing character designs, video footage, or storyboard ideas. In reference generation, multiple images can define people or objects, and video can guide movement; the prompt then explains what should happen in the new scene. This workflow is particularly useful for brand stories, episodic short dramas, and product videos in which the same character appears across multiple clips. Rather than describing everything through text alone, creators can provide subject information separately from shot requirements, reducing the need to rewrite the setup for every attempt. Official materials also describe storyboard controls for specifying each shot's duration, framing, viewpoint, and camera movement, helping organize complex clips around the intended narrative. Prepare clear references that do not conflict and explain the purpose of each one. Review character details, sound, and transitions segment by segment, especially when reusing characters across different scenes.
- Reference-Based Generation
- Character Consistency
Key features
Multiple image and video references
Define characters or scenes with images and optionally add reference video to guide movement and style.
Character continuity
Carry subject features from references into new scenes, making it suitable for recurring-character content.
Segmented storytelling and audio
Use multiple prompts to organize scenes and, depending on the reference video, retain its original audio or generate new sound.
