Kling 4.0 Omni Reference
Kling Omni Reference
Give Kling 4.0 pictures, clips and a voice to follow — a character, a product, a camera move, a way of speaking — and say in the prompt what each one is for.
- Up to 15 files
- 10 images · 5 clips · 3 voices
- Type @ to name them
- Stereo sound
- 3–30 s
Omni Reference examples
Real Kling 4.0 renders from this site, with the exact prompt. Use one as a starting point.
“Handheld tracking shot: a young woman in a yellow raincoat runs across a rain-soaked Tokyo crosswalk at night, neon signs reflecting in puddles, she laughs and looks back at the camera, rain hitting umbrellas, distant traffic and city hum”
Text · Kling 4.0 · 720P · 5s
Use this prompt“A street violinist plays on a cobblestone bridge in Prague at golden hour, slow push-in from a low angle, pigeons take off behind her, the violin melody echoes over the river”
Text · Kling 4.0 · 720P · 5s
Use this prompt“Aerial drone shot following a lone motorcyclist across red desert dunes at sunrise, long shadows, dust trailing behind, the engine roars and wind rushes past”
Text · Kling 4.0 · 1080P · 5s
Use this prompt“Two old friends sit at a sunny café window. The man smiles and says: "You still take it with no sugar?" The woman laughs and answers: "Some things never change." Gentle camera drift, cups clinking, soft café music”
Text · Kling 4.0 · 720P · 10s
Use this prompt“Overhead close-up of a bowl of ramen as chopsticks lift glossy noodles, steam curling up, a soft-boiled egg splits open, slurping and kitchen chatter in the background”
Text · Kling 4.0 · 720P · 5s
Use this prompt“Cinematic wide shot of a lighthouse on a cliff during a storm, huge waves crash against the rocks, the light beam sweeps through the rain, thunder rumbles”
Text · Kling 4.0 · 720P · 5s
Use this prompt
Guide the video with references
Omni Reference is how Kling 4.0 follows your files without using them as frames: a character's face and outfit, a product's shape and logo, the layout of a scene, the rhythm of a camera move, the voice of a speaker. It grew out of Kling O1 and Kling 3.0 Omni, which took far fewer references at once.
Each upload is numbered in order — @Image1, @Image2, @Video1, @Audio1. Type @ in the prompt to insert one, then give it a job: "@Image1 is the main character, keep her face and coat. Follow the camera path of @Video1." Without a job, the model guesses.
Clips are 3–30 seconds and 30 seconds in total; voice clips 5–30 seconds. Reference images are free. With a reference video, you pay the output seconds plus the longer of the output or the video, at 14 credits a second at 720P and 20 at 1080P.
Kling is a trademark of Kuaishou Technology. Kling V4 AI is an independent service and is not affiliated with or endorsed by Kuaishou.
How to use Kling 4.0 online
Three steps in the browser: no GPU, no app, no subscription.
- 01
Describe the shot
Say what happens: subject, action, camera, light and sound. Or start from an example.
- 02
Add pictures or references
Optional: a first (and last) frame, or reference images, clips and a voice sample to follow.
- 03
Generate and download
Pick the resolution, length and shape, see the price, and download the MP4 when it's ready.
Omni Reference questions
Choose Reference above, add images, clips or a voice sample, then describe the new video and what each file is for. Kling 4.0 keeps what you point it to and generates the rest.
By upload order: @Image1, @Image2, @Video1, @Audio1. Type @ in the prompt box to pick one, or use the chips under the box.
Up to 10 images, 5 video clips and 3 voice clips, 15 files in total. Clips are 3–30 seconds each and 30 seconds in total; voice clips are 5–30 seconds and need at least one image or clip. Kling's app also takes saved elements; this site doesn't.
Reference images are free. A reference video changes the bill: you pay the output seconds plus the longer of the output or the video, at 14 credits a second at 720P (20 at 1080P). A 10-second 720P clip with a 6-second reference video is (10 + 10) × 14 = 280 credits. The button shows the total.
No. Kling O1 was Kuaishou's first unified multimodal video model (December 2025), later upgraded to Kling 3.0 Omni. Omni Reference in Kling 4.0 is its successor and takes more files per generation: up to 15.
Image to video uses your picture as the actual first (and last) frame. Reference mode uses pictures and clips as guides, so the video can show the character or product in a brand-new shot.
They're used only to make your video and deleted 30 days later. Use files you own or have permission to use.
Try Kling 4.0 now — your first video is free
Write a sentence, pick a length, and hear it come to life.
Start creating