Omni references
Combine up to 10 images, 5 videos and 3 audio clips, with a maximum of 15 reference files in one task. Use images for appearance, videos for movement, and audio for voice guidance. Add audio references in Canvas.
Kling 4.0
Generate reference-guided video with Kling 4.0. Bring characters, scenes, and camera direction together with native audio in your Arcloop project.
Kling 4.0 is Kuaishou’s video model. Arcloop currently supports reference-to-video: combine visual references with a prompt, describe how the scene unfolds, and refine the result in Canvas.
Kling 4.0
About Kling 4.0
Kling 4.0 is Kuaishou’s video generation model, developed around realistic performance, controlled movement and cinematic storytelling. Its improvements address both quiet character moments and demanding action: facial expressions, continuous movement and camera direction can work together to shape a scene.
The model’s Omni Reference approach combines visual and audio guidance for character identity, performance, camera movement and narrative rhythm. Images can establish a person, product or location; video references can communicate movement and pacing. This supports workflows such as short dramas, action sequences and product storytelling.
Arcloop currently connects Kling 4.0 Reference. Add an image or video in the creation box, describe the action, camera and sound, then continue into your project’s Canvas. Review the available model settings and credit estimate before generating, and use the result as the starting point for your next revision.
Kling 4.0 Video Prompt Examples
A five-second continuous live-action shot on the exposed deck of a storm-battered ocean liner. Begin close to rain-soaked hands playing a black grand piano, water trembling across the polished surface. The camera steadily pulls back to reveal an adult pianist in a soaked dark formal suit, still playing with quiet concentration as the deck lists beneath him. Behind the piano, a wave strikes the railing and throws white spray across the background. Cold overcast light, wet wood, heavy fabric moving in the wind. A sparse piano melody remains audible beneath wind, rain, and the impact of the sea.
A five-second live-action scene with exactly two shots. Two adult former acquaintances unexpectedly recognize each other on a rainy railway platform at night. Shot 1, 0–2.5 seconds: over the woman's softly blurred shoulder, a medium close-up of the man. Recognition interrupts his expression, then a restrained smile appears. Shot 2, 2.5–5 seconds: reverse angle over the man's shoulder, a medium close-up of the woman. She pauses, meets his gaze, and gives a small answering nod. Match their eyelines and maintain the 180-degree axis. Warm platform lamps, cool rain beyond the canopy, natural skin texture. Continuous rain and distant train ambience, no dialogue or music.
Scene description: Two men struggle beside a parked sedan on a rain-soaked city street. Follow their movement around the car, then move closer as a battered man opens a black briefcase. Overcast light, wet pavement, brick buildings and restrained handheld action photography. This describes the supplied clip; the original generation prompt is unavailable.
A five-second continuous live-action military science-fiction shot inside a severely cramped, damaged combat dropship plunging toward a war-torn planet. Film from a camera wedged against the left cabin wall, looking perpendicular across the narrow aisle: the lead space infantryman is seen in strong right-facing side profile, his enormous battered charcoal power armor, thick shoulder plates and enclosed helmet almost filling the frame. A compact rifle is secured against his chest. Low overhead pipes, scratched metal ribs, dangling webbing and two armored soldiers squeezed onto wall seats make the limited space unmistakable. During the first two seconds a violent impact throws the cabin sideways; he braces his boots and grabs an overhead strap while sparks spit from a broken panel and red emergency strobes cut through thin smoke. The side hatch on screen-right jolts open, exposing fast-moving cloud, distant anti-aircraft flashes and a burning landscape rising beneath the descending craft. A green bailout lamp flashes. He releases his harness, takes one forceful step and jumps through the hatch to screen-right in the final beat, emergency jump pack firing briefly outside. Keep the camera inside the cabin, preserving the side-on view; short impact jolts, no orbit or angle change. Heavy industrial realism, practical armor surfaces, cold daylight against red alarms. Audio: hull impacts, rattling metal, urgent warning siren, armored boots and explosive wind through the hatch.
A single continuous 5-second gothic fantasy film shot. A regal adult vampire countess in an ornate black velvet cloak stands beside a tall broken stained-glass window in an ancient castle library at midnight. Moonlight catches her pale face. She slowly opens a centuries-old silver locket, looks toward the camera, and reveals a restrained knowing smile with small elegant fangs. A few bats glide silently past the window in the distant background while loose pages lift in a cold draft. Slow cinematic dolly-in from a medium portrait to a close portrait, sophisticated practical costume details, natural skin, candlelight against cool moonbeams, mist outside the window. Soft wind, candle crackle and a low mysterious cello note. Atmospheric and romantic, no attack, no blood, no gore, no jump scare, no text or watermark.
Kling 4.0 capabilities
Camera and performance
Direct a close-up, tracking shot or shot/reverse-shot exchange. Describe a clear action, expression and camera move, with separate beats when a scene needs a cut.
Native audio
Generate dialogue, ambience and effects with the picture. Officially listed dialogue languages include Chinese, English, Japanese, Korean, Spanish, Portuguese, German, French and Hindi. Specify the language, speaker and line in your prompt; audio is always generated.
How to use Kling 4.0
Prepare your shot on this page, then review settings and generate in your project’s Canvas.
- 01
Prepare your prompt and references
Add image or video references in the creation box above. Describe the subject, action, camera movement and sound you want.
- 02
Continue in Canvas
Click Generate Now. Sign in if prompted; your prompt and selected references will be brought into a project in Canvas.
- 03
Review settings and generate
In Canvas, review the selected model, references, output options and credit estimate, then submit your generation.
- 04
Preview and refine
Preview the completed result in your project. Adjust the prompt or references and generate another version when needed.
What you can create
Short dramas
Reunions, confrontations and quiet conversations driven by character and emotion.
Films and trailers
Opening scenes, dramatic reveals and trailer moments that set the mood.
Animation
Expressive characters and imagined worlds, from everyday stories to science fiction.
Tips for Kling 4.0
Practical guidance for writing and refining prompts. Review each result against the shot you intended.
- 01
Keep prompts focused: 300–500 words
Aim for 300–500 words in an English prompt. Prioritize the subject, key action, camera movement and sound. Too many details, repeated descriptions or competing instructions can dilute the main idea and make it harder for the model to identify what matters. Keep each shot focused on one main action and one camera move.
- 02
Give each reference a job
Say which image defines the character and which video guides movement. Reuse the same appearance and wardrobe references across shots, then check faces after turns and wide camera moves.
- 03
Write dialogue as performance
Name the speaker, specify the language and put the exact line in quotes. Start with short turns; describe the listener’s reaction so two characters do not speak at once.
- 04
Direct the soundscape
List the sounds you want, such as rain, boots and cabin alarms. State your music preference explicitly, then review the result: unwanted music or imperfect lip-sync can still occur.
- 05
Use negative prompts selectively
Add short exclusions to the same prompt when needed, such as “no subtitles” or “no background music.” Avoid long lists of negatives. For visual corrections, describe the desired result directly, such as “the character remains in side profile.”
Kling 4.0 Pricing
Text / image references
Generation settings: 5 sec · 720p
Estimated credits: 785 credits / 5 sec
Text / image references
Generation settings: 5 sec · 1080p
Estimated credits: 1,175 credits / 5 sec
With 5 sec reference video
Generation settings: 5 sec · 720p
Estimated credits: 940 credits / 5 sec
With 5 sec reference video
Generation settings: 5 sec · 1080p
Estimated credits: 1,410 credits / 5 sec
Prices shown are estimates. Actual charges are determined by final settlement.
Kling 4.0 alternatives
Choose an alternative by the inputs and shot you want to create. Settings below describe the current Arcloop integrations.
Kling 4.0
Reference-led character performance, deliberate camera movement and short-drama or action scenes.
- Duration
- 3–30 seconds
- Reference limits
- 10 images / 5 videos / 3 audio; 15 total
- Output resolution
- 720p / 1080p
Seedance 2.5
Larger reference sets, connected story beats and films assembled from multiple shots.
- Duration
- 4–30 seconds
- Reference limits
- 30 images / 10 videos / 10 audio
- Output resolution
- 480p / 720p / 1080p
Wan 3.0
Text-to-video, first-frame animation and mixed-reference shots, with optional audio.
- Duration
- 2–30 seconds
- Reference limits
- 10 images / 5 videos / 5 audio
- Output resolution
- 480p / 720p / 1080p
Kling 4.0 · FAQ
Which Kling 4.0 model does Arcloop use?
Arcloop offers Kling 4.0 Reference for multi-reference video creation. Add your references, describe the scene, and continue in Canvas.
How do I add references?
Add image or video references in the quick composer. Continue in Canvas for the complete model settings and supported reference types.
How much does a generation cost?
Check the current credit estimate in the creation workspace before submitting. Resolution, duration, and references can affect the final cost.
Can I create video with sound?
Yes. The connected reference model generates audio by default. Describe the sound, dialogue, or atmosphere you want in the prompt.
Start creating with Kling 4.0
Generate reference-guided video with Kling 4.0. Bring characters, scenes, and camera direction together with native audio in your Arcloop project.