A character can look convincing in a sketch and become difficult to read as soon as it moves. The outline changes, a hand disappears against the background, or a camera move pulls attention away from the action. Planning an animation means deciding which visual qualities must survive those changes.

For animators, Kling 4.0 brings together keyframes, shared references, stylized rendering and generated audio as distinct ways to direct a scene. The Kling 4.0 overview introduces these controls through specifications and examples.



Keyframes: shape a 3–30-second character action

Kling 4.0 supports clips from 3 to 30 seconds and up to 10 keyframe images. Keyframes give an action visual landmarks: the pose before movement, the important gesture and the reaction afterward. For an animator, this makes the model relevant to blocking a scene around decisions already made in drawings.

For an illustrative messenger scene, prepare four poses: pause beneath a streetlamp, reach into a pocket, unfold a note, and look up in surprise. Keep the note visible and the hands clear against the coat in each drawing. Choose an intended clip length that gives the reveal and reaction room to register.

Use the same medium framing across these key poses so the action, rather than a changing camera, carries the scene. Additional keyframes can clarify a difficult gesture or an important transition. Each image should add an essential state of the action; filling all 10 slots is not a creative goal.

Shared references: define character, motion and style

The reference system supports up to 15 combined assets, with per-type limits of 10 images, five videos totaling 30 seconds, and seven subjects. These allowances share the 15-asset ceiling. Their value for animation is the ability to communicate different parts of the production brief through different kinds of input.

Assign the character drawings to appearance: proportions, the triangular blue coat and the yellow scarf. Use a short, authorized movement reference to show the pace of the hand reaching into the pocket. A lighting study can establish warm lamplight against deep blue shadows, while a texture reference describes the felt surface.

References establish the visual and movement vocabulary; keyframes establish the action's changing poses. Give each selected reference one clear job, and resolve contradictory designs before assembling the pack. Use your own drawings and footage or material cleared for this use, including any identifiable person's likeness.

Style controls: build a felt, clay or 3D scene

Kling 4.0's published style examples include felt, clay and 3D looks. For the messenger, felt suggests visible fibers, soft surface detail and simple shapes; a clay treatment would call for a different material description. Name the desired rendering and its important traits instead of relying on a broad label such as cinematic.

Connect the style request to the reference pack and key poses. The coat's triangular silhouette and yellow scarf are essential design features, while the brick texture behind the character can be flexible. Put those priorities in the prompt so material, lighting and motion directions serve the same design.

Illustrative prompt combining the visual and audio controls:

Felt-art animation of a small messenger in a triangular blue coat and yellow scarf beneath one streetlamp at night. In a fixed medium shot, she takes a folded note from her pocket, opens it and looks up. Warm lamplight, deep blue shadows and visible fiber texture. Keep the coat silhouette, scarf color and note consistent. Soft footsteps and paper rustle over quiet street ambience. She whispers, "Someone changed the plan."

Pair this scene prompt with the four key poses and the chosen appearance, movement and texture references. The prompt is an example to evaluate, rather than a report of generated footage. A fixed camera and one speaker keep the visual gesture and the spoken reaction easy to follow.

Generated audio: connect voices and scene sounds

Kling 4.0 supports audiovisual generation, stereo output and improved lip-sync for dialogue and performances. For animation, these controls make sound part of scene direction: a footstep establishes contact, paper rustle marks the note opening, and a short spoken line gives the reaction a second channel.

Specify which sounds belong to which visible actions. In the messenger scene, keep street ambience low beneath the footsteps, let the paper sound coincide with unfolding, and place the spoken line after the character looks up. Show the mouth clearly during the line and leave enough time for the expression to settle.

Stereo can give the surroundings a sense of space, but the main action should remain understandable when the clip is played through a phone speaker. Review the timing of the gesture, paper sound and mouth movement together. A visually strong scene still needs a sound balance that lets the important moment land.

Review the action, then refine one element

Check the key poses against the clip, then inspect the coat silhouette, scarf color, hands and note across transitions. Record a specific time and change, such as "00:03: the scarf turns orange." Keep this acceptance pass short and tied to the original references rather than judging only the overall atmosphere.

The described targeted-editing and omniref controls provide another route for revisions guided by text and references. For example, request a darker background or a more surprised expression while asking to retain the character design and action. Compare the revised candidate with the same key poses, reference sheet and audio cues.