How to Prompt Kling 4.0: Camera, Motion and Defaults to Kill
Prompt structure that works and the music, subtitle, and automatic cut defaults you must write against to maintain creative control in Kling 4.0.
Key takeaways
- Kling 4.0 applies three aggressive defaults by default: background music, burnt-in subtitles, and automatic shot cuts.
- Prompts must include explicit negative blocks to suppress music, soundtrack scores, and subtitles before execution.
- Short shots follow the official four-block template, while identity-critical shots use an expanded thirteen-block structure.
- Middle scenes must be loaded into image references as @Image1 through @ImageN, with every transition wipe scripted.
- Keyframes mode cannot combine with Omni Reference in the same job; composition and face locking require separate runs.
Suppressing the three default behaviors in Kling 4.0
The official Kling 4.0 guide documents three defaults that the model applies unless the prompt blocks them: it adds background music, it burns in subtitles, and its multi-shot Smart Storyboard cuts on its own with no toggle. Our prompts block all three with 'SFX only, no music, no score, no song, no soundtrack', 'No subtitles, no captions, no watermarks, no UI' and 'One continuous shot, no cuts, no angle changes, no scene changes'.
When deploying commercial prompts to Kling 4.0, unmanaged default behaviors degrade production quality instantly. Left unguided, the model's audio synthesis engine populates video exports with generic ambient music tracks, while the vision decoder burns in distracting subtitle text across lower-third pixels. Simultaneously, the automated Smart Storyboard feature introduces autonomous camera cuts, fracturing continuous narrative takes into disjointed montages. To establish baseline creative control, every studio prompt must begin with strict negative declarations that silence music scores, eliminate on-screen text, and enforce continuous single-take camera choreography.
The guide says music and subtitles are the only negatives truly worth writing. We phrase everything else positively, because the guide warns that negatives can increase the likelihood of errors.
While silencing audio and graphical watermarks requires explicit negative phrasing, applying broad negative constraints to visual elements frequently degrades generation quality. As emphasized in official documentation, overusing negative tokens confuses the attention mechanism and increases the likelihood of compositional defects. Outside of audio and subtitle suppression, our prompt engineering relies exclusively on affirmative descriptive directives.
Prompt architecture: four-block templates and house styles
Structure we use for short shots is the official four-block Omni template: Asset Roles, Overall Settings, a timecoded Sequence of Events, then Audio. For identity-critical shots or anything over about 8 seconds we switch to a numbered house style: scene context, active references, location map, first frame, optics, camera, action timing, physics, lighting, audio, character acting, style, and a quality block that lists what must not change.
Related
Writing Camera Moves for AI Video Models
Naming the move, one move per shot, what models do with compound moves: camera mechanics, dialogue constraints, and previz rules.
How to Prompt Seedance 2.5 for Dialogue Scenes
Two synthetic characters, two voices, two lines — and the right mouth moved for each on the first try. The structure that made it work, and the failures that taught us each rule.
Kling 4.0 vs Seedance 2.5: Cost per Second and What Each Gets Wrong
Kling 4.0 really is less than half the price of Seedance 2.5 on the same platform. What the price comparison leaves out is the three default behaviours you have to prompt against, and the question nobody has answered yet: faces.