Kling AI VIDEO 3.0 / 3.0 Omni

Kling AI VIDEO 3.0 / 3.0 Omni (Kuaishou) is a video engine within the FilmRadar stack. Prompt structure: Subject -> Action -> Scene -> Camera -> Lighting & Mood.

Character Lock

Clearly define core subjects at the START of the prompt; maintain consistent clothing/hairstyle/props/role across shots while changing scene/camera/action. Utilize Element Reference / Video Element Reference (Omni) + clean reference images. Keep character descriptions verbatim.

Strongest Keywords

• close-up
• wide shot
• low angle
• slow push-in
• pan
• tilt
• tracking shot
• dolly-in
• orbit
• extreme close-up
• medium close-up
• full body shot

Template

[Shot 1: {framing}] {subject}, {action}, {scene}. {camera_move}. {lighting_mood}.
[Shot 2: {framing}] {same subject verbatim}, {action}. {camera_move}.
[Shot 3: {framing}] {same subject verbatim}, {action}. {camera_move}.
High consistency, {style}, {aspect_ratio}.

Origin

Grounded from official documentation. Doc: https://kling.ai/blog/kling-ai-prompt-guide. kling.ai/blog Prompt Guide (2026-07-03) + Video-3.0-User-Guide, fetched today. Structure/multishot/audio/languages taken literally from the official source.