Real outputs from the public examples of this model on Replicate, with the prompt and settings that produced them. They were not generated live on this page.
Prompt
A Wes Anderson style movie trailer with symmetrical framing and pastel colors. Opening: a perfectly centered shot of a grand pink hotel facade with teal shutters. A bellboy in a pillbox hat pushes a luggage cart precisely down the center of a hallway. Quick lateral tracking shots: a mustachioed concierge stamps a passport, twin sisters in matching yellow dresses ride a tandem bicycle through a village square, a detective examines a painting through a magnifying glass while standing in an absurdly tiny elevator. Every shot is perfectly balanced and color-coordinated in pastels. Final symmetrical shot of the entire cast standing in a row on the hotel steps. Text in a vintage serif font: "KLING V3 OMNI VIDEO" then "NOW ON REPLICATE". Whimsical harpsichord music, the sound of a bell ding.
A lavish period drama movie trailer set in 18th century Versailles. Opening: a sweeping crane shot over manicured gardens with fountains at golden hour. Cut to a grand ballroom filled with dancers in elaborate silk gowns and powdered wigs, candlelight from enormous chandeliers. A mysterious woman in a crimson dress locks eyes with a nobleman across the room. Quick cuts: secret letters being passed, a duel at dawn in the mist, tears on a beautiful face. Final shot: the woman stands alone on a balcony overlooking Paris at sunset. Elegant gold text fades in: "KLING V3 OMNI VIDEO" then "NOW ON REPLICATE". Sweeping orchestral strings, harpsichord notes, dramatic crescendo.
Kuaishou Technology (founded 2011 by Su Hua and Cheng Yixiao) operates one of China's two dominant short-video platforms and runs the KLING AI lab. Following the success of Kling 1.x and Kling v3, the team released Kling v3 Omni in 2025 as the multimodal-conditioning variant of v3. Omni accepts not only text and a single image but also audio reference, video reference, character-reference images, sketches, depth maps and pose skeletons, making it the broadest conditioning surface among Chinese video models. Omni is exposed through klingai.com, the Kling Studio editor and the Kling API.
Diffusion Transformer (DiT) with multimodal conditioning adapters
Kling v3 Omni shares the same core DiT denoiser as Kling v3 but adds a suite of conditioning adapters that inject heterogeneous control signals into the diffusion backbone. The model accepts text prompts (bilingual LLM encoder), first frame, end frame, subject-reference images (for locked identity), pose / skeleton videos (for motion transfer), depth maps (for 3D-aware control), sketches (for layout), audio reference (for music-driven motion and lip-sync) and short video references (for style transfer). Each modality is encoded by a dedicated adapter network and merged via cross-attention layers into the 3D spatio-temporal latent stream. Joint audio-video diffusion produces synchronized sound, including lip-sync to a provided dialogue audio track. Training uses paired multimodal-conditioning data including motion-capture libraries, sketch-to-video pairs and music videos.
Text-, image-, pose-, sketch-, depth-, audio- and video-reference conditioning
Lip-sync to a user-provided dialogue audio track
Motion transfer from a pose / skeleton video
Subject-reference for locked character identity
Sketch- and depth-controlled layout for storyboards
1080p / 30 fps, native synchronized audio
Available via Kling Studio and API
Strongest conditioning surface among Chinese video models
Best for: animation studios, lip-sync content, motion-capture-driven workflows.
Entraînement et licence
Closed corpus including platform video, licensed footage, motion-capture libraries, sketch / depth annotation pairs and music videos with dense multimodal captions.
Licence: Proprietary commercial licence via Kling AI / Kuaishou terms.
curl https://railwail.com/api/v1/videos/generations \
-H "Authorization: Bearer $RAILWAIL_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "kling-v3-omni",
"prompt": "A slow drone shot over a misty pine forest at sunrise"
}'
# Response: {"job_id": "...", "status": "queued", ...}
# Poll until status is completed, failed or cancelled:
curl https://railwail.com/api/v1/jobs/JOB_ID \
-H "Authorization: Bearer $RAILWAIL_API_KEY"
import os
import time
import requests
API = "https://railwail.com/api/v1"
headers = {"Authorization": f"Bearer {os.environ['RAILWAIL_API_KEY']}"}
job = requests.post(
f"{API}/videos/generations",
headers=headers,
json={
"model": "kling-v3-omni",
"prompt": "A slow drone shot over a misty pine forest at sunrise",
},
).json()
while True:
status = requests.get(f"{API}/jobs/{job['job_id']}", headers=headers).json()
if status["status"] in ("completed", "failed", "cancelled"):
break
time.sleep(5)
print(status["status"], status.get("output_url"))
const API = "https://railwail.com/api/v1";
const headers = {
Authorization: `Bearer ${process.env.RAILWAIL_API_KEY}`,
"Content-Type": "application/json",
};
const job = await fetch(`${API}/videos/generations`, {
method: "POST",
headers,
body: JSON.stringify({
model: "kling-v3-omni",
prompt: "A slow drone shot over a misty pine forest at sunrise"
}),
}).then((r) => r.json());
let status;
do {
await new Promise((r) => setTimeout(r, 5000));
status = await fetch(`${API}/jobs/${job.job_id}`, { headers }).then((r) => r.json());
} while (!["completed", "failed", "cancelled"].includes(status.status));
console.log(status.status, status.output_url);
