Kling 4.0 Reference to Video API
Reference to Video with Kling 4.0 by Kuaishou. Kling's newest video model — up to 30 s with stereo sound on every clip.
per second of video · 720p
Operational99.92% uptime (30 days)median 6m 44sTry Kling 4.0 Reference to Video
Up to 10 images for characters, products, style or scene. Refer to them in the prompt as @Image1, @Image2…
Up to 5 MP4/MOV clips for motion, camera moves or effects (3–30 s each, 30 s in total, ≤ 200 MB each). Refer to them as @Video1, @Video2… With a reference video, the output seconds plus the longer of the output and the reference videos are billed at a lower per-second rate.
Output frame shape.
Output resolution. Higher resolutions cost more per second.
Length of the generated clip in seconds (3–30s). You pay per second.
Create a free account with Google or email — pay only for what you generate.
Your video will appear here
Write a prompt, pick your settings and hit Run. You only pay for successful generations.
Also available: Kling 4.0 Text to Video · Kling 4.0 Image to Video
See the Kling 4.0 examples and prompts
Kling 4.0 Reference to Video API
/api/v1/run/kuaishou/kling-4.0/reference-to-videoSubmits an asynchronous job and returns 202 with a generation id. Poll GET /api/v1/generations/{id} or pass webhook_url. See job lifecycle.
Input parameters
| Name | Type | Description | Affects price |
|---|---|---|---|
promptrequired | string | Describe the shot: subject, action, camera movement, lighting and style. | No |
reference_image_urls | string[] (URLs) | Up to 10 images for characters, products, style or scene. Refer to them in the prompt as @Image1, @Image2… Max items: 10 | No |
reference_video_urls | string[] (video URLs) | Up to 5 MP4/MOV clips for motion, camera moves or effects (3–30 s each, 30 s in total, ≤ 200 MB each). Refer to them as @Video1, @Video2… With a reference video, the output seconds plus the longer of the output and the reference videos are billed at a lower per-second rate. Max items: 5 | Yes |
aspect_ratio | string (enum) | Output frame shape. Values: 16:9 · 9:16 · 1:1 · 21:9 Default: 16:9 | Yes |
resolution | string (enum) | Output resolution. Higher resolutions cost more per second. Values: 720p · 1080p Default: 720p | Yes |
duration | integer | Length of the generated clip in seconds (3–30s). You pay per second. Range: 3–30s Default: 5 | Yes |
webhook_url | string (URL) | Optional. We POST the finished generation here, signed with your webhook secret. Webhooks | No |
Request
const API = "https://cheapergen.com/api/v1";
const headers = {
Authorization: `Bearer ${process.env.CHEAPERGEN_API_KEY}`,
"Content-Type": "application/json",
};
// 1. Submit the job
let job = await fetch(`${API}/run/kuaishou/kling-4.0/reference-to-video`, {
method: "POST",
headers,
body: JSON.stringify({
"prompt": "@Image1 walks along the beach at sunset, following the camera move of @Video1",
"reference_image_urls": [
"https://cdn.cheapergen.com/samples/reference-1.jpg"
],
"reference_video_urls": [
"https://cdn.cheapergen.com/samples/reference-1.mp4"
],
"aspect_ratio": "16:9",
"resolution": "720p",
"duration": 5
}),
}).then((r) => r.json());
// 2. Poll until it finishes (or pass webhook_url and skip polling)
while (job.status === "queued" || job.status === "running") {
await new Promise((r) => setTimeout(r, 3000));
job = await fetch(`${API}/generations/${job.id}`, { headers }).then((r) => r.json());
}
console.log(job.status === "succeeded" ? job.output.video.url : job.error);Response
{
"id": "gen_8c2QmX1nq0k4",
"object": "generation",
"model": "kuaishou/kling-4.0",
"endpoint": "kuaishou/kling-4.0/reference-to-video",
"mode": "reference-to-video",
"status": "queued",
"input": {
"prompt": "@Image1 walks along the beach at sunset, following the camera move of @Video1",
"reference_image_urls": [
"https://cdn.cheapergen.com/samples/reference-1.jpg"
],
"reference_video_urls": [
"https://cdn.cheapergen.com/samples/reference-1.mp4"
],
"aspect_ratio": "16:9",
"resolution": "720p",
"duration": 5
},
"output": null,
"error": null,
"cost": {
"amount_usd": 0.2255,
"currency": "usd",
"estimated": true,
"official_usd": null,
"list_price_source": "official",
"list_price_label": "Official",
"saved_usd": null
},
"source": "api",
"started_at": null,
"completed_at": null,
"urls": {
"get": "https://cheapergen.com/api/v1/generations/gen_8c2QmX1nq0k4",
"cancel": "https://cheapergen.com/api/v1/generations/gen_8c2QmX1nq0k4/cancel"
},
"created_at": "2026-10-15T09:21:04Z"
}Kling 4.0 pricing
Pay per second of video as you go, with no subscription required (monthly plans are optional). Failed generations are refunded automatically.
| Tier | Official price | CheaperGen | You save |
|---|---|---|---|
720p audio included | $0.1486/s | −5% | |
1080p audio included | $0.2226/s | −5% |
- Official price: Kling developer API list price for Kling 4.0, kling.ai/dev/pricing (1 unit = $0.157 / s at 720p, 1.5 units = $0.235 / s at 1080p; with input video $0.094 / $0.141 per second of output plus the longer of the output and the input video; audio always on). Prices verified October 18, 2026.
- Audio is always generated and included in the price.
- With a reference video, the output seconds plus the longer of the output and the reference videos are billed at a lower per-second rate.
More from Kuaishou
Popular Kuaishou models, on the same API key.

Kling3.0 Omni
Text · Image · Reference → Video
Kling 3.0 with multi-image references for consistent characters

Kling3.0
Text · Image → Video
Kling's flagship video model — cinematic motion, native audio, up to 15 s

Kling3.0 Turbo
Text · Image → Video
Faster Kling 3.0 with native audio on every clip
Similar video models
Popular text-to-video models from other developers.

Seedance2.5
Text to Video · Image to Video · Reference to Video · Video Edit
ByteDance's newest video model — longer, sharper, more controllable

MiniMaxH3
Text · Image · Reference → Video
Hailuo H3 — expressive motion and physics, up to 2K

MiniMaxH3 Fast
Text · Image · Reference → Video
MiniMax H3 in seconds, not minutes — an accelerated build of the open weights

Seedance2.0 Fast
Text · Image · Reference → Video
Seedance 2.0 quality at lower latency and cost