BLACK FRIDAY

الجمعة السوداء: خصم 50% على السنوي، و30% رصيد إضافي مع Pro

ينتهي خلال --ي --:--:--احصل على خصم 50%

How to Use Kling 4.0: References, Keyframes, Camera and Sound

How to use Kling 4.0 with Omni Reference, keyframes, camera direction and sound prompts—plus a beginner workflow from first idea to review.

How to Use Kling 4.0: References, Keyframes, Camera and Sound

Short answer: To use Kling 4.0, pick a starting point (text, image, or references), confirm the model and settings in the studio, then write one clear shot. For asset-led control, use Omni Reference with @ tags; for timed visual moments, use Keyframes. Add timestamped camera moves and a separate [Audio] section for dialogue, ambience and effects.

Last verified: 2026-10-21

Choose your starting point

If you have an idea but no image, start with text to video. If the look of the subject matters, begin with a photograph or illustration and use image to video. If you need to guide several moments, prepare references and keyframes for a model interface that supports them.

Want a character image to copy the movement in a reference video? Start with Kling Motion Control.

Pick one small goal for the first attempt: a camera move around a product, a person turning toward the camera, or a short environmental scene. A focused shot is easier to judge than a complete advertisement.

To use Kling AI image to video, start with one image you have permission to use. Check the face, hands or product outline at full size. A blurred edge or an object hidden behind another object leaves the model less visual information to work with. Choose a clear original image rather than a screenshot with interface elements.

Prepare the crop for the final placement before uploading to Kling 4.0 image to video. Leave space in the direction the subject will move. For a person turning, keep the head and shoulders visible; for an object sliding, show its supporting surface. If the selected model offers different image-input modes, check whether your image sets the opening frame or serves as a reference. They guide different parts of the result.

Check the actual model

Open Kling 4.0 and check the model selector in the studio. Use the exact model name and settings shown for your task. Confirm the image-input mode, available duration, aspect ratio, resolution and audio options before submitting. Do not assume that controls described for one version are included in another. Check the displayed credit estimate, and save the selected settings with your prompt so the next attempt has a clear comparison point.

Choosing between Kling 4.0 and Kling 4.0 Flash? Compare length, resolution and frame control in our Kling 4.0 Flash guide.

Write one clear shot

Start with the subject, setting and action. Then add a camera instruction and the details that should stay consistent. “A woman turns toward the window as the camera slowly moves closer” is a workable direction. “Make it epic, cinematic, viral and perfect” does not explain what should happen.

For an image input, describe what changes over time. Give the subject and camera separate instructions: the subject turns slightly toward the window; the camera stays in the same position. Once that version is usable, you can try one slow camera move. “Locked camera” is a direction to review in the output, not a guarantee.

State the amount of movement and the relationships that should hold. A small head turn asks for less change than a full spin. If a hand holds a cup, specify that the grip stays on the handle; if a bottle sits on a table, keep its base in contact with the surface. Avoid asking a tightly cropped subject to reveal large unseen areas in the first attempt.

For example: “The seated person turns their head slightly toward the window, then pauses. Their hands remain resting on the table. The camera stays fixed in a medium shot.” Use the Kling 4.0 prompts guide for other scene structures.

Set duration and framing

Choose an aspect ratio for the intended placement. Landscape is useful for a wide scene; portrait can suit a mobile feed. Keep the important subject away from edges that may be cropped later.

Match the action to the clip length. A short duration suits a single reveal. A longer clip can hold several beats, but each beat should lead naturally into the next. Start with a simple version before increasing complexity.

Need a longer clip? See how to extend Kling 4.0 videos up to 2 minutes.

Review before you publish

Watch the complete clip, then inspect moments where the subject turns, objects overlap or the camera changes direction. Check faces, fingers, product labels, geometry, unwanted cuts and any spoken words.

  • If almost nothing moves, name one visible action and its direction. Check that the source crop gives it room. Keep the image and settings unchanged for the next comparison.
  • If a face, hand or product bends during motion, reduce the movement first. If the problem is already visible in the source image, repair or replace that input before another generation.
  • If the camera drifts or an unwanted cut appears, remove competing camera directions and simplify the change from the starting pose. Review the transition, not just the opening and final frames.
  • If upload or submission fails before a clip exists, note the exact error and task status. Check the file requirements shown in the studio and keep the task ID for support. A failed submission needs a different check from an unwanted visual result.

Refine one decision at a time

Keep the useful parts of the first brief. Change a single variable, such as camera speed or the main action, and compare the next result. This gives you a clearer explanation of why the output improved or got worse.

My Creations keeps your prompts, selected model, settings and task results together so you can review a generation and plan the next version. Download files you want to keep within the 14-day storage window.

Download a candidate clip and review the actual file in your editing workflow. Check its dimensions, audio and the full sequence at the size viewers will see. A promising preview may still contain a distracting frame, unreadable label or awkward ending. Keep the original download as well as your edited version.

Trim unusable opening or closing frames, then add accurate captions, product text and any licensed music in an external editor. If you crop for a mobile feed, check the subject again after cropping. Export for the intended placement and play that final file before publishing. Keep the final shot only if it delivers the one goal you set.

Sources: Kling AI Image-to-Video Guide; Kling AI Image-to-Video Motion Prompts. These references explain image-led motion prompting; the controls available in the selected studio model may differ.

Open the studio

Explore the model


To use Kling 4.0, choose Omni Reference for asset-led creation or Keyframes for precisely timed visual moments. In Omni Reference, assign each asset a role with @ tags; in Keyframes, place images along the timeline. Write camera moves as timestamped instructions, then describe dialogue, ambience, music, and effects in a separate [Audio] section. Keep appearance references, action instructions, and sound directions distinct so each has a clear purpose.

Last verified: 2026-10-21

Kling 4.0 Usage Guide: Omni Reference vs Keyframes

Kling 4.0 offers two entry points. Omni Reference is the default and most comprehensive, supporting text-to-video and keyframe-to-video alongside reference-driven work. Keyframes concentrates on reproducing predefined visual moments. Choose according to what needs the strongest control.

ComparisonOmni ReferenceKeyframes
Best forAn idea alone, or image and video assetsPredefined visual moments requiring precise reproduction
What it controlsAppearance, style, actions, camera movementsDistribution of key moments along the timeline; pixel-level consistency in first-frame and first-and-last-frame workflows
UploadsImages, videos, voice profilesKeyframe images only
Aspect ratio behaviour21:9, 16:9, 1:1, 9:16, AutoWith images, follows the first keyframe; without images, 21:9, 16:9, 1:1, 9:16
Can combine?Cannot combine the two entry points in one taskCannot combine the two entry points in one task

For a first reference-based project in the Kling 4.0 video generator, use Omni Reference when you need to explain what each supplied asset contributes.

Reference images and assets

Kling 4.0 reference prompting works best when every upload has a specific job. An image might define a character, composition, or storyboard moment; a video might supply action pacing or a camera path. For multiple characters, identify each separately and describe their interaction.

  • Text: up to 8,000 tokens.
  • Images: up to 10, each up to 30 MB, with aspect ratios from 2:5 to 5:2.
  • Videos: up to 5, each 3–30 seconds and up to 300 MB; combined duration is limited to 30 seconds.
  • Voice profiles: up to 3 per task, sourced from 5–30-second audio files up to 50 MB each.
  • Elements: up to 7, including up to 3 video elements with 3–8-second sources.
  • Total: 15 reference assets. Reference videos, editing videos, and video elements share the 5-clip and 30-second quota.

These limits apply together. Before adding another asset, check both its individual allowance and the shared totals. More references help only when their roles remain clear.

@Image1: preserve the character's appearance.
@Image2: reference the room's composition and materials.
@Video1: reference the action pacing.
Scene and texture: a quiet studio with textured plaster walls; visual style: natural daylight; camera tone: observant.
0-3 s: The character reaches toward a ceramic cup; the camera slowly pushes in.
3-7 s: The character lifts the cup and smiles; the camera tracks gently right.
[Audio] Soft room ambience and a ceramic clink; no background music; no subtitles.

Use the Kling 4.0 Omni Reference guide for reference workflows and the character consistency guide when appearance preservation is central.

Keyframes

Kling 4.0 accepts up to 10 keyframe images. You can define a first frame, a last frame, both endpoints, or middle frames at timeline positions. This makes Keyframes useful when a gesture, entrance, or composition must occur at a particular moment.

For example, place a raised-hand image at 3 seconds and a turned-around pose at 7 seconds. Another sequence could show character B entering through a door at 6 seconds. For a weather transition, place a darkened sky at 5 seconds and rain at 10 seconds.

Describe the movement connecting those images. For an extreme-wide-to-close-up progression, specify the timestamp of the close-up and the camera movement leading into it. Uploaded images determine the aspect ratio through the first keyframe, so prepare that image around the intended composition.

The Keyframes entry accepts no reference assets. Keep reference-driven tasks in Omni Reference, and use the Kling 4.0 keyframes guide for timeline planning. Kling 4.0 duration runs from 3–30 seconds in one-second steps, with a 10-second default; see the Kling 4.0 video length guide.

Camera direction

Give each camera instruction a subject, direction, and time window. Use a push-in to approach a detail, a pull-out to reveal surroundings, a pan to turn across a scene, tracking to follow movement, and an orbit to move around a subject. Kling 4.0 supports these movements and multi-shot direction.

0-3 s: Push in toward the parked car's front grille.
3-6 s: Track alongside the car as it moves down the street.
6-9 s: Orbit around the car as it comes to a stop.
9-12 s: Pull out to reveal the surrounding square.

When a reference clip defines the movement, write: “Reference the camera path and pacing of @Video1. Replace the element with the car in @Image1 and the background with @Image2.” This separates camera behaviour from the objects appearing in the reference.

Prefer visible details over “premium look”: describe side backlighting outlining edges and a blurred background. Explore motion control and Kling prompts for further direction examples.

Sound

Kling 4.0 supports two-channel stereo and more accurate lip sync. Supported languages include Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French, and Hindi. Chinese dialect examples include Sichuanese and Cantonese; English accents include American, British, and Indian.

Audio uploads serve only as voice profile references, with up to 3 profiles per task using the file limits above. Music and sound effects cannot be uploaded; describe them in the audio section for generation from the visuals. Audio-visual sync and Smart Storyboard are on by default, with no toggles.

[Audio]
Dialogue: The character says in British English, "I saved you a seat."
Ambient sounds: Quiet cafe conversation surrounding the speakers.
Sound effects: A cup touches its saucer as the character sits.
No background music.
No subtitles.

State the speaker, exact words, and desired language or accent. Describe sound events alongside their visible causes. Use descriptive instructions generally, with “no background music” and “no subtitles” explicitly included when unwanted.

FAQ

Can I combine Omni Reference and Keyframes?

No. The two entry points cannot be combined in one task, although Omni Reference itself supports keyframe-to-video.

Why does my keyframe sequence use a different aspect ratio?

With uploaded images, Kling 4.0 follows the first keyframe's aspect ratio. Check that image before preparing the remaining frames.

Can I upload a soundtrack?

No. Audio input is limited to voice profile references. Describe background music and effects in the [Audio] section.

Does Kling 4.0 Flash support multiple keyframes?

Kling 4.0 Flash supports up to 20 seconds and a maximum of 720p, without start-and-last frames or multiple keyframes.

What should I change when references are followed poorly?

Clarify each asset's role and replace abstract descriptions. Start with one sentence covering element, scene, action, and emotion; then add camera, timing, audio, and preservation instructions.

Sources

Try Kling AI Today

Build your Kling 4.0 prompt around clear asset roles, timed movement, and a dedicated audio section.

Start creating