Kling Omni Video O3
kwaivgi/kling-video-o3/reference-to-video
Kling Omni Video O3 Reference-to-Video generates creative videos using character, prop, or scene references from multiple viewpoints. Extracts subject features and creates new video content while maintaining identity consistency across frames. Supports audio generation. Ready-to-use REST API, best performance, no cold starts, affordable pricing.
Examples
Parameters
| Name | Type | Default | Constraints | Description |
|---|---|---|---|---|
| promptPrompt | textarea | — | ≤ 5000 chars | The positive prompt for the generation. |
| imagesImages | image_upload | — | image/* · 1–7 items | Reference images. With a reference video: image elements ≤ 4; without a reference video: ≤ 7 |
| resolutionGeneration Mode | select | 720P | 720P | 1080P | |
| aspect_ratioAspect Ratio | select | 16:9 | 16:9 | 9:16 | 1:1 | The aspect ratio of the generated video. |
| durationDuration | select | 5 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | … | The duration of the generated media in seconds (3-15). |
| soundSound | boolean | false | — | Whether to generate audio for the video. |
| keep_original_soundKeep Original Sound | boolean | true | — | Whether to keep the original sound from the reference video. |
| multi_shotMulti Shot | boolean | — | — | Whether to generate multi-shot video When true: the prompt parameter is invalid. When false: the shot_type and multi_prompt parameters are invalid |
| shot_typeShot Type | select | — | customize | intelligence | Shot type for the generation. |
| multi_promptMulti Prompt | array<object> | — | — | List of multi-prompt elements for the generation. |
| ↳durationDuration | number | 5 | — | The duration of this shot in seconds. |
| ↳promptPrompt | text | — | — | The prompt for this shot. |
API
Call this model through one unified REST API. Get a key on the API Keys page.
cURL
# 1) Submit — returns { "task_uuid": "..." }
curl -X POST "https://www.namifusion.com/api/v1/marketplace/run/kwaivgi/kling-video-o3/reference-to-video" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"input": {
"prompt": "Use reference image 1 as the female character and reference.",
"images": [
"https://assets-public.namifusion.com/marketplace/images/2026-04-14/54456cddb7de.png"
],
"resolution": "720P",
"aspect_ratio": "16:9",
"duration": 5,
"sound": false,
"keep_original_sound": false,
"shot_type": "intelligence"
}
}'
# 2) Poll until status is "completed", then read the output URLs
curl "https://www.namifusion.com/api/v1/marketplace/run/tasks/TASK_UUID" \
-H "Authorization: Bearer YOUR_API_KEY"Python
import time, requests
API_KEY = "YOUR_API_KEY"
HEADERS = {"Authorization": f"Bearer {API_KEY}"}
# 1) Submit
resp = requests.post(
"https://www.namifusion.com/api/v1/marketplace/run/kwaivgi/kling-video-o3/reference-to-video",
headers=HEADERS,
json={
"input": {
"prompt": "Use reference image 1 as the female character and reference.",
"images": [
"https://assets-public.namifusion.com/marketplace/images/2026-04-14/54456cddb7de.png"
],
"resolution": "720P",
"aspect_ratio": "16:9",
"duration": 5,
"sound": False,
"keep_original_sound": False,
"shot_type": "intelligence"
}
},
)
resp.raise_for_status() # 401/402/429/5xx stop here instead of polling a bad task
task = resp.json()
# 2) Poll until a terminal state (completed / failed / cancelled).
# This model is allowed up to 300s server-side.
deadline = time.time() + 360
while task.get("status") not in ("completed", "failed", "cancelled"):
if time.time() > deadline:
raise TimeoutError(f"still {task.get('status')} — keep the task_uuid and poll later")
time.sleep(3)
poll = requests.get(f"https://www.namifusion.com/api/v1/marketplace/run/tasks/{task['task_uuid']}", headers=HEADERS)
poll.raise_for_status()
task = poll.json()
print(task["status"], task.get("output"))JavaScript
const API_KEY = "YOUR_API_KEY";
const HEADERS = { Authorization: `Bearer ${API_KEY}` };
// 1) Submit
const resp = await fetch("https://www.namifusion.com/api/v1/marketplace/run/kwaivgi/kling-video-o3/reference-to-video", {
method: "POST",
headers: { ...HEADERS, "Content-Type": "application/json" },
body: JSON.stringify({
"input": {
"prompt": "Use reference image 1 as the female character and reference.",
"images": [
"https://assets-public.namifusion.com/marketplace/images/2026-04-14/54456cddb7de.png"
],
"resolution": "720P",
"aspect_ratio": "16:9",
"duration": 5,
"sound": false,
"keep_original_sound": false,
"shot_type": "intelligence"
}
}),
});
if (!resp.ok) throw new Error(`submit failed: ${resp.status} ${await resp.text()}`);
let task = await resp.json();
// 2) Poll until a terminal state (completed / failed / cancelled).
// This model is allowed up to 300s server-side.
const deadline = Date.now() + 360 * 1000;
while (!["completed", "failed", "cancelled"].includes(task.status)) {
if (Date.now() > deadline) throw new Error(`still ${task.status} — keep the task_uuid and poll later`);
await new Promise((r) => setTimeout(r, 3000));
const poll = await fetch(`https://www.namifusion.com/api/v1/marketplace/run/tasks/${task.task_uuid}`, { headers: HEADERS });
if (!poll.ok) throw new Error(`poll failed: ${poll.status}`);
task = await poll.json();
}
console.log(task.status, task.output);Documentation
Kling Video O3 Reference-to-Video
Effortless Character-Consistent Video Generation
Kling Omni Video O3 Reference-to-Video enables developers to create high-quality videos using reference images and optional reference videos. This model ensures identity consistency across frames, supports flexible aspect ratios and durations, and offers optional audio generation. With no cold starts and affordable pricing, it’s optimized for seamless integration into creative workflows.
🚀 Key Features
- Character-Consistent Generation: Maintains the identity of characters and elements throughout the video.
- Multi-Reference Support: Combine multiple reference images to create coherent scenes.
- Optional Reference Video: Use a reference video for motion guidance, style transfer, or scene continuity.
- Flexible Output: Supports multiple aspect ratios (16:9, 9:16, 1:1) and durations ranging from 3 to 15 seconds.
- Sound Options: Keep original audio or generate synchronized sound effects.
- Multi-Prompt Support: Guide scene transitions and narrative flow with chained prompt segments.
🛠️ Technical Specifications
| Parameter | Type | Required | Default | Range/Options | Description |
|---|---|---|---|---|---|
| prompt | string | Yes | - | - | Text description of the desired scene and action |
| video | string | No | - | - | Reference video URL for motion or style guidance |
| images | array | No | [] | ≤4 with video, ≤7 without video | Reference images of characters, elements, or styles |
| keep_original_sound | boolean | No | True | - | Retain original sound from the reference video |
| sound | boolean | No | False | - | Generate synchronized audio for the video |
| aspect_ratio | string | No | 16:9 | 16:9, 9:16, 1:1 | Aspect ratio of the generated video |
| duration | integer | No | 5 | 3–15 | Length of the generated video in seconds |
| shot_type | string | No | customize | customize, intelligent | Editing mode for generation |
| multi_prompt | array | No | - | - | Additional prompt segments for scene transitions |
💰 Pricing
| Attribute | 720P | 1080P |
|---|---|---|
| No Ref Video + Silent | 0.0840 | 0.1120 |
| No Ref Video + With Sound | 0.1120 | 0.1400 |
| With Ref Video + Silent | 0.1260 | 0.1680 |
| With Ref Video + With Sound | 0.1540 | 0.1960 |
Billing Rules:
- All prices are in USD.
- Rates are determined by resolution, reference video usage, and audio status.
- Supported duration: 3–15 seconds.
💡 Best Use Cases
- Character-Driven Storytelling: Create scenes starring specific characters from reference images.
- Social Media Content: Produce personalized short-form videos for platforms like TikTok and Instagram.
- Marketing & Ads: Generate spokesperson or brand ambassador videos from still photos.
- Creative Concepting: Combine multiple characters or elements into new scenarios for rapid ideation.
- Style Transfer: Use a reference video to guide motion and visual style for new content.
🔗 Related Models
- Kling Video O3 Image-to-Video: Animate a single image into video.
- Kling Video O3 Text-to-Video: Generate videos from text prompts.
- Kling Video O3 Video Edit: Edit existing videos with natural-language instructions.
Related models
VT Lip Sync
VT Lip Sync synchronizes mouth movements for a translated video by using the translated target video, translated audio, translated subtitle timing, and the original source video with its original subtitle timing.
Kling Omni Video O3
Kling Omni Video O3 Reference-to-Video generates creative videos using character, prop, or scene references from multiple viewpoints. Extracts subject features and creates new video content while maintaining identity consistency across frames. Supports audio generation. Ready-to-use REST API, best performance, no cold starts, affordable pricing.
Kling Omni Video O1 Reference-to-Video
Kling Omni Video O1 Reference-to-Video generates creative videos using character, prop, or scene references from multiple viewpoints. Extracts subject features and creates new video content while maintaining identity consistency across frames. Ready-to-use REST API, best performance, no cold starts, affordable pricing.
Kling Omni Video O1 Reference-to-Video
Kling Omni Video O1 Reference-to-Video generates creative videos using character, prop, or scene references from multiple viewpoints. Extracts subject features and creates new video content while maintaining identity consistency across frames. Ready-to-use REST API, best performance, no cold starts, affordable pricing.
Kling 3.0 Standard
Kling 3.0 Standard Motion Control transfers motion from reference videos to animate still images. Upload a character image and a motion clip (dance, action, gesture), and the model extracts the movement to generate smooth, realistic video. Ready-to-use REST inference API, best performance, no cold starts, affordable pricing.
seedance-2-0 reference-to-video
The "Reference-to-Video" feature of Seedance 2.0 is the ultimate solution for visual stylistic unity. It precisely extracts artistic styles, lighting tones, or compositional intents from reference materials and seamlessly integrates them into newly generated videos, ensuring a highly consistent visual language for your creative series.
Alibaba WAN 2.6
Alibaba WAN 2.6 Reference-to-Video seamlessly transforms character, prop, or scene references—supporting both single and multi-view inputs—into high-quality video sequences. It excels at preserving identity, style, and layout while delivering fluid, coherent motion. Experience peak performance via our production-ready REST API, featuring zero cold starts and cost-effective pricing.