SkyReels V4 Series

POST

tasks

SkyReels V4 Series

curl --request POST \
  --url https://api.qingbo.ai/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "duration": 123,
  "resolution": "<string>",
  "aspect_ratio": "<string>",
  "video_urls": [
    "<string>"
  ],
  "first_frame_image": "<string>",
  "end_frame_image": "<string>",
  "mid_frame_images": [
    {}
  ],
  "ref_images": [
    {}
  ],
  "prompt_optimizer": true
}
'

import requests

url = "https://api.qingbo.ai/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "duration": 123,
    "resolution": "<string>",
    "aspect_ratio": "<string>",
    "video_urls": ["<string>"],
    "first_frame_image": "<string>",
    "end_frame_image": "<string>",
    "mid_frame_images": [{}],
    "ref_images": [{}],
    "prompt_optimizer": True
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: '<string>',
    action: '<string>',
    prompt: '<string>',
    duration: 123,
    resolution: '<string>',
    aspect_ratio: '<string>',
    video_urls: ['<string>'],
    first_frame_image: '<string>',
    end_frame_image: '<string>',
    mid_frame_images: [{}],
    ref_images: [{}],
    prompt_optimizer: true
  })
};

fetch('https://api.qingbo.ai/v1/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.qingbo.ai/v1/tasks"

	payload := strings.NewReader("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"duration\": 123,\n  \"resolution\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"video_urls\": [\n    \"<string>\"\n  ],\n  \"first_frame_image\": \"<string>\",\n  \"end_frame_image\": \"<string>\",\n  \"mid_frame_images\": [\n    {}\n  ],\n  \"ref_images\": [\n    {}\n  ],\n  \"prompt_optimizer\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.qingbo.ai/v1/tasks")
  .header("Content-Type", "application/json")
  .body("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"duration\": 123,\n  \"resolution\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"video_urls\": [\n    \"<string>\"\n  ],\n  \"first_frame_image\": \"<string>\",\n  \"end_frame_image\": \"<string>\",\n  \"mid_frame_images\": [\n    {}\n  ],\n  \"ref_images\": [\n    {}\n  ],\n  \"prompt_optimizer\": true\n}")
  .asString();

SkyReels V4 is a unified multimodal video-and-audio generation model that jointly produces synchronized picture and sound from text, image and video inputs, at 480p to 1080p and 3 to 15 seconds, and can extend an existing video. This page covers skyreels-v4-fast and skyreels-v4-std: both tiers expose the same actions, parameters and value ranges — Fast favors speed, Std favors fidelity — so switching tiers means changing model.

Quick start

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: skyreels-demo-001" \
  -d '{
    "model": "skyreels-v4-fast",
    "action": "generate",
    "prompt": "A shiba inu in a spacesuit walking on the moon, the camera pushing in slowly",
    "duration": 5,
    "resolution": "1080p",
    "aspect_ratio": "16:9"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "skyreels-v4-std",
    "action": "generate",
    "prompt": "She walks toward the camera, hair moving in the wind",
    "first_frame_image": "https://cdn.example.com/first.jpg",
    "end_frame_image": "https://cdn.example.com/last.jpg",
    "duration": 8,
    "resolution": "1080p"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "skyreels-v4-fast",
    "action": "generate",
    "prompt": "Follow the camera motion in @video_1, but set it on a snowy mountain",
    "video_urls": ["https://cdn.example.com/source.mp4"],
    "resolution": "1080p"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "skyreels-v4-fast",
    "action": "extend",
    "prompt": "Extend @video_1: someone walks over and sits on the sofa",
    "video_urls": ["https://cdn.example.com/source.mp4"],
    "duration": 8,
    "resolution": "720p"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "skyreels-v4-std",
    "action": "generate",
    "prompt": "@Actor-1 walks through a neon-lit street at night, the camera tracking sideways",
    "ref_images": [
      {
        "tag": "@Actor-1",
        "type": "image",
        "image_urls": ["https://cdn.example.com/actor-a.jpg"]
      }
    ],
    "duration": 6,
    "resolution": "1080p"
  }'

A successful submission returns a task_id. Retrieve the result with GET /v1/tasks/{task_id}, wait inline with Prefer: wait, or configure a webhook.

Request parameters

string

required

Either skyreels-v4-fast or skyreels-v4-std.

string

default:"generate"

generate — generate a video. The mode follows the media fields you send: prompt alone is text-to-video; first_frame_image, end_frame_image or mid_frame_images is image-to-video; ref_images or video_urls is reference-driven generation. Image-to-video fields and reference fields cannot be combined.
extend — video extension. Continues the one source video in video_urls for duration seconds.

string

required

The video description. When you send video_urls, ref_images or mid_frame_images, the prompt must contain the matching @ tag; under action: "extend" reference @video_1.

integer

Video length in seconds, an integer from 3 to 15.

action: "generate" with video_urls: the output follows the reference video (3 seconds when shorter) and this parameter has no effect
action: "extend": the length of the extension

string

default:"1080p"

Output resolution: 480p, 720p or 1080p; defaults to 1080p.

string

Frame ratio: 16:9, 9:16, 1:1, 4:3 or 3:4. Ignored when a first/end frame image or a reference video is present, and under action: "extend" — the ratio follows the input material.

string[]

Video URLs, at most 1, MP4 / MOV. The Nth entry gets the tag @video_N, which the prompt must reference (for example Follow the camera motion in @video_1); without it the upstream rejects the request and nothing is billed.

action: "generate": motion and subject reference, 1 to 10 seconds
action: "extend": required, the source video to extend, 1 to 15 seconds

string

Starting frame image URL (jpg / jpeg / png / gif / bmp). action: "generate" only.

string

Closing frame image URL; combine with the first frame for start/end control. action: "generate" only.

object[]

Mid keyframe list, up to 6 entries, available in the image-to-video mode of action: "generate" only. Each entry holds:

tag: a tag starting with @ that must appear in the prompt
image_url: the image URL
time_stamp: the moment it appears, in seconds. -1 (the default) leaves it unspecified; otherwise 0 < time_stamp < duration

object[]

Reference image list, available in the reference mode of action: "generate" only and mutually exclusive with the image-to-video fields. All entries must share the same type. Each entry holds:

tag: a tag starting with @ that must appear in the prompt
type: image for an ordinary reference image, grid for a single image composed of several tiles
image_urls: an array of image URLs
audio_url: a voice audio URL, allowed only when type is image, up to 15 seconds

boolean

default:"true"

Whether to auto-optimize the prompt.

See Submit Task for callback_url, callback_events, Prefer: wait, Idempotency-Key and the maximum-cost header. Any other parameter returns 400 and is not billed.

Limits

Condition	Limit
`action: "generate"` with `video_urls`	With reference video: duration follows the reference (max 10s)
`action: "extend"`	Extension continues exactly one source video (MP4/MOV, at most 15 s) for the requested duration (3-15 s); images, first/end frames and reference images cannot be sent, and the aspect ratio follows the source video.
Every request	This model produces one result per task; n is not supported

Item	Limit
Duration `duration`	3 to 15 seconds; under `generate` with `video_urls`, follows the reference video, capped at 10 seconds
Resolution `resolution`	`480p` / `720p` / `1080p`, default `1080p`
Video `video_urls`	1 clip, MP4 / MOV; 1 to 10 seconds under `generate`, 1 to 15 seconds under `extend`
Mid keyframes `mid_frame_images`	Up to 6 entries
Reference images `ref_images`	`type: image`: 1 to 3 entries with 1 to 5 images each; `type: grid`: 1 entry with 1 image
Mutually exclusive	`first_frame_image` / `end_frame_image` / `mid_frame_images` cannot be combined with `ref_images` / `video_urls`

All input material must be publicly downloadable over HTTP(S).

Pricing

Billing is output resolution tier × seconds; without resolution the 1080p tier applies. Every skyreels-v4-std tier sits above the matching skyreels-v4-fast tier.

Request	Rate tier	Billed seconds
`generate` without `video_urls`	Ordinary tier	The requested `duration`
`generate` with `video_urls`	Reference-video tier	The reference video’s length (3 seconds when shorter); the requested `duration` has no effect
`extend`	Reference-video tier	The requested `duration`

The reference-video tier is above the ordinary tier at the same resolution. The input video’s own seconds are not billed; reference images (ref_images), keyframe images and prompt length carry no separate charge.

Per-model rates are in price_config on GET /v1/models, and in the console’s Model Market.What a single call actually cost is the cost field on the task response — an integer quota at 500,000 quota = 1 USD.

You are billed only for a delivered video. Failed and cancelled tasks, and upstream successes that carry no deliverable video, are refunded in full. The final charge is the cost field on the task response — an integer quota, not dollars (500,000 quota = 1 USD).

Response

Completed task

{
  "task_id": "task-wave1775285160b950328499",
  "model": "skyreels-v4-fast",
  "action": "generate",
  "status": "completed",
  "progress": "100%",
  "created_at": 1775285160,
  "completed_at": 1775285320,
  "result": {
    "videos": [
      {
        "expires_at": 1775371560,
        "url": ["https://cdn.example.com/result.mp4"]
      }
    ]
  },
  "billing_status": "settled",
  "cost": 61875,
  "urls": {
    "get": "https://api.qingbo.ai/v1/tasks/task-wave1775285160b950328499",
    "cancel": "https://api.qingbo.ai/v1/tasks/task-wave1775285160b950328499/cancel"
  }
}

Generated videos are returned in result.videos, with picture and audio in the same file; url is an array and expires_at is the link expiry in Unix seconds, so copy the file to your own storage before then. See Query Task Status for the full field reference.

Available models

Model ID	Resolution	Duration (s)	Action	Notes
`skyreels-v4-fast`	`480p` / `720p` / `1080p`	Any integer 3 to 15	`generate` `extend`	Speed first
`skyreels-v4-std`	`480p` / `720p` / `1080p`	Any integer 3 to 15	`generate` `extend`	Fidelity first

SkyReels V4 Series

curl --request POST \
  --url https://api.qingbo.ai/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "duration": 123,
  "resolution": "<string>",
  "aspect_ratio": "<string>",
  "video_urls": [
    "<string>"
  ],
  "first_frame_image": "<string>",
  "end_frame_image": "<string>",
  "mid_frame_images": [
    {}
  ],
  "ref_images": [
    {}
  ],
  "prompt_optimizer": true
}
'

import requests

url = "https://api.qingbo.ai/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "duration": 123,
    "resolution": "<string>",
    "aspect_ratio": "<string>",
    "video_urls": ["<string>"],
    "first_frame_image": "<string>",
    "end_frame_image": "<string>",
    "mid_frame_images": [{}],
    "ref_images": [{}],
    "prompt_optimizer": True
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: '<string>',
    action: '<string>',
    prompt: '<string>',
    duration: 123,
    resolution: '<string>',
    aspect_ratio: '<string>',
    video_urls: ['<string>'],
    first_frame_image: '<string>',
    end_frame_image: '<string>',
    mid_frame_images: [{}],
    ref_images: [{}],
    prompt_optimizer: true
  })
};

fetch('https://api.qingbo.ai/v1/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.qingbo.ai/v1/tasks"

	payload := strings.NewReader("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"duration\": 123,\n  \"resolution\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"video_urls\": [\n    \"<string>\"\n  ],\n  \"first_frame_image\": \"<string>\",\n  \"end_frame_image\": \"<string>\",\n  \"mid_frame_images\": [\n    {}\n  ],\n  \"ref_images\": [\n    {}\n  ],\n  \"prompt_optimizer\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.qingbo.ai/v1/tasks")
  .header("Content-Type", "application/json")
  .body("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"duration\": 123,\n  \"resolution\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"video_urls\": [\n    \"<string>\"\n  ],\n  \"first_frame_image\": \"<string>\",\n  \"end_frame_image\": \"<string>\",\n  \"mid_frame_images\": [\n    {}\n  ],\n  \"ref_images\": [\n    {}\n  ],\n  \"prompt_optimizer\": true\n}")
  .asString();

Text Generation

Image Generation

Video Generation

Audio

Tools & Embeddings

Task Management

Account & Usage

SkyReels V4 Series

Quick start

Request parameters

Limits

Pricing

Response

Available models

​Quick start

​Request parameters

​Limits

​Pricing

​Response

​Available models

​Related

Quick start

Request parameters

Limits

Pricing

Response

Available models

Related