> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apimart.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# SkyReels V4 视频生成

>  - 提供 Fast（速度优先）和 Std（质量优先）两档模型
- 支持文生视频（T2V）、图生视频（I2V）、多模态参考（Omni）三种模式，由传参自动路由
- 支持 480p / 720p / 1080p 分辨率，3 ~ 15 秒时长
- 支持首帧/尾帧/关键帧、参考图、参考视频、拼贴图、视频扩展、声纹同步等高阶能力
- 异步处理模式，返回任务 ID 用于后续查询 

<RequestExample>
  ```bash cURL theme={null}
  curl --request POST \
    --url https://api.apimart.ai/v1/videos/generations \
    --header 'Authorization: Bearer <token>' \
    --header 'Content-Type: application/json' \
    --data '{
      "model": "skyreels-v4-fast",
      "prompt": "A serene forest at sunset with golden light filtering through the trees.",
      "duration": 5,
      "resolution": "1080p",
      "aspect_ratio": "16:9",
      "prompt_optimizer": true
    }'
  ```

  ```python Python theme={null}
  import requests

  url = "https://api.apimart.ai/v1/videos/generations"

  payload = {
      "model": "skyreels-v4-fast",
      "prompt": "A serene forest at sunset with golden light filtering through the trees.",
      "duration": 5,
      "resolution": "1080p",
      "aspect_ratio": "16:9",
      "prompt_optimizer": True
  }

  headers = {
      "Authorization": "Bearer <token>",
      "Content-Type": "application/json"
  }

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

  print(response.json())
  ```

  ```javascript JavaScript theme={null}
  const url = "https://api.apimart.ai/v1/videos/generations";

  const payload = {
    model: "skyreels-v4-fast",
    prompt: "A serene forest at sunset with golden light filtering through the trees.",
    duration: 5,
    resolution: "1080p",
    aspect_ratio: "16:9",
    prompt_optimizer: true
  };

  const headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
  };

  fetch(url, {
    method: "POST",
    headers: headers,
    body: JSON.stringify(payload)
  })
    .then(response => response.json())
    .then(data => console.log(data))
    .catch(error => console.error('Error:', error));
  ```

  ```go Go theme={null}
  package main

  import (
      "bytes"
      "encoding/json"
      "fmt"
      "io/ioutil"
      "net/http"
  )

  func main() {
      url := "https://api.apimart.ai/v1/videos/generations"

      payload := map[string]interface{}{
          "model":            "skyreels-v4-fast",
          "prompt":           "A serene forest at sunset with golden light filtering through the trees.",
          "duration":         5,
          "resolution":       "1080p",
          "aspect_ratio":     "16:9",
          "prompt_optimizer": true,
      }

      jsonData, _ := json.Marshal(payload)

      req, _ := http.NewRequest("POST", url, bytes.NewBuffer(jsonData))
      req.Header.Set("Authorization", "Bearer <token>")
      req.Header.Set("Content-Type", "application/json")

      client := &http.Client{}
      resp, err := client.Do(req)
      if err != nil {
          panic(err)
      }
      defer resp.Body.Close()

      body, _ := ioutil.ReadAll(resp.Body)
      fmt.Println(string(body))
  }
  ```

  ```java Java theme={null}
  import java.net.http.HttpClient;
  import java.net.http.HttpRequest;
  import java.net.http.HttpResponse;
  import java.net.URI;

  public class Main {
      public static void main(String[] args) throws Exception {
          String url = "https://api.apimart.ai/v1/videos/generations";

          String payload = """
          {
            "model": "skyreels-v4-fast",
            "prompt": "A serene forest at sunset with golden light filtering through the trees.",
            "duration": 5,
            "resolution": "1080p",
            "aspect_ratio": "16:9",
            "prompt_optimizer": true
          }
          """;

          HttpClient client = HttpClient.newHttpClient();
          HttpRequest request = HttpRequest.newBuilder()
              .uri(URI.create(url))
              .header("Authorization", "Bearer <token>")
              .header("Content-Type", "application/json")
              .POST(HttpRequest.BodyPublishers.ofString(payload))
              .build();

          HttpResponse<String> response = client.send(request,
              HttpResponse.BodyHandlers.ofString());

          System.out.println(response.body());
      }
  }
  ```

  ```php PHP theme={null}
  <?php

  $url = "https://api.apimart.ai/v1/videos/generations";

  $payload = [
      "model" => "skyreels-v4-fast",
      "prompt" => "A serene forest at sunset with golden light filtering through the trees.",
      "duration" => 5,
      "resolution" => "1080p",
      "aspect_ratio" => "16:9",
      "prompt_optimizer" => true
  ];

  $ch = curl_init($url);
  curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
  curl_setopt($ch, CURLOPT_POST, true);
  curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($payload));
  curl_setopt($ch, CURLOPT_HTTPHEADER, [
      "Authorization: Bearer <token>",
      "Content-Type: application/json"
  ]);

  $response = curl_exec($ch);
  curl_close($ch);

  echo $response;
  ?>
  ```

  ```ruby Ruby theme={null}
  require 'net/http'
  require 'json'
  require 'uri'

  url = URI("https://api.apimart.ai/v1/videos/generations")

  payload = {
    model: "skyreels-v4-fast",
    prompt: "A serene forest at sunset with golden light filtering through the trees.",
    duration: 5,
    resolution: "1080p",
    aspect_ratio: "16:9",
    prompt_optimizer: true
  }

  http = Net::HTTP.new(url.host, url.port)
  http.use_ssl = true

  request = Net::HTTP::Post.new(url)
  request["Authorization"] = "Bearer <token>"
  request["Content-Type"] = "application/json"
  request.body = payload.to_json

  response = http.request(request)
  puts response.body
  ```

  ```swift Swift theme={null}
  import Foundation

  let url = URL(string: "https://api.apimart.ai/v1/videos/generations")!

  let payload: [String: Any] = [
      "model": "skyreels-v4-fast",
      "prompt": "A serene forest at sunset with golden light filtering through the trees.",
      "duration": 5,
      "resolution": "1080p",
      "aspect_ratio": "16:9",
      "prompt_optimizer": true
  ]

  var request = URLRequest(url: url)
  request.httpMethod = "POST"
  request.setValue("Bearer <token>", forHTTPHeaderField: "Authorization")
  request.setValue("application/json", forHTTPHeaderField: "Content-Type")
  request.httpBody = try? JSONSerialization.data(withJSONObject: payload)

  let task = URLSession.shared.dataTask(with: request) { data, response, error in
      if let error = error {
          print("Error: \(error)")
          return
      }

      if let data = data, let responseString = String(data: data, encoding: .utf8) {
          print(responseString)
      }
  }

  task.resume()
  ```

  ```csharp C# theme={null}
  using System;
  using System.Net.Http;
  using System.Text;
  using System.Threading.Tasks;

  class Program
  {
      static async Task Main(string[] args)
      {
          var url = "https://api.apimart.ai/v1/videos/generations";

          var payload = @"{
              ""model"": ""skyreels-v4-fast"",
              ""prompt"": ""A serene forest at sunset with golden light filtering through the trees."",
              ""duration"": 5,
              ""resolution"": ""1080p"",
              ""aspect_ratio"": ""16:9"",
              ""prompt_optimizer"": true
          }";

          using var client = new HttpClient();
          client.DefaultRequestHeaders.Add("Authorization", "Bearer <token>");

          var content = new StringContent(payload, Encoding.UTF8, "application/json");
          var response = await client.PostAsync(url, content);
          var result = await response.Content.ReadAsStringAsync();

          Console.WriteLine(result);
      }
  }
  ```
</RequestExample>

<ResponseExample>
  ```json 200 theme={null}
  {
    "code": 200,
    "data": [
      {
        "status": "submitted",
        "task_id": "task_01KPEY5H3NQ2W8D7T6VB3F9GR4"
      }
    ]
  }
  ```

  ```json 400 theme={null}
  {
    "error": {
      "code": 400,
      "message": "请求参数无效",
      "type": "invalid_request_error"
    }
  }
  ```

  ```json 401 theme={null}
  {
    "error": {
      "code": 401,
      "message": "身份验证失败，请检查您的API密钥",
      "type": "authentication_error"
    }
  }
  ```

  ```json 402 theme={null}
  {
    "error": {
      "code": 402,
      "message": "账户余额不足，请充值后再试",
      "type": "payment_required"
    }
  }
  ```

  ```json 422 theme={null}
  {
    "error": {
      "code": 422,
      "message": "参数互斥或取值非法（例如同时传入 I2V 与 Omni 字段）",
      "type": "invalid_request_error"
    }
  }
  ```

  ```json 429 theme={null}
  {
    "error": {
      "code": 429,
      "message": "请求过于频繁，请稍后再试",
      "type": "rate_limit_error"
    }
  }
  ```

  ```json 500 theme={null}
  {
    "error": {
      "code": 500,
      "message": "服务器内部错误，请稍后重试",
      "type": "server_error"
    }
  }
  ```
</ResponseExample>

## 认证

<ParamField header="Authorization" type="string" required>
  所有接口均需要使用 Bearer Token 进行认证

  获取 API Key：

  访问 [API Key 管理页面](https://apimart.ai/keys) 获取您的 API Key

  使用时在请求头中添加：

  ```
  Authorization: Bearer YOUR_API_KEY
  ```
</ParamField>

## 生成模式

SkyReels V4 通过请求字段自动路由到对应模式，**无需指定 `mode` 字段**：

| 模式              | 触发条件                                                               | 能力                      |
| --------------- | ------------------------------------------------------------------ | ----------------------- |
| **T2V**（文生视频）   | 只传 `prompt` 及通用字段                                                  | 纯文本驱动生成                 |
| **I2V**（图生视频）   | 传入 `first_frame_image` / `end_frame_image` / `mid_frame_images` 任一 | 首帧、尾帧、关键帧控制             |
| **Omni**（多模态参考） | 传入 `ref_images` / `ref_videos` 任一                                  | 主体参考、拼贴图、动作参考、视频扩展、声纹同步 |

<Warning>
  **严格互斥**：I2V 字段（`first_frame_image` / `end_frame_image` / `mid_frame_images`）与 Omni 字段（`ref_images` / `ref_videos`）不能同时出现，否则返回 422。
</Warning>

<Note>
  **`@tag` 机制**：当使用 `mid_frame_images` / `ref_images` / `ref_videos` 时，每个元素需声明一个以 `@` 开头的 `tag`（如 `@image1`、`@Actor-1`、`@video1`），并且 `tag` **必须出现在 `prompt` 中**。

  可以把 `prompt` 理解为"剧本"，`tag` 则是指向具体素材（图片 / 视频）的"角色指针"—— 例如 prompt 写 `"@Actor-1 走进 @video1 的场景"`，系统据此将 `@Actor-1` 对应的参考图主体和 `@video1` 对应的动作参考注入到生成过程。
</Note>

## 请求参数

### 通用字段

<ParamField body="model" type="string" required>
  支持以下两个档位：

  | 模型                 | 定位                       | 适用场景            |
  | ------------------ | ------------------------ | --------------- |
  | `skyreels-v4-fast` | 速度优先                     | 快速预览、批量生成、日常内容  |
  | `skyreels-v4-std`  | 质量优先（价格较 Fast 高 25\~30%） | 关键画面、细节要求高、正式交付 |

  <Warning>
    **必须显式传递 `model` 字段，无默认值。**
  </Warning>

  <Tip>
    **计费与分辨率、是否使用 `ref_videos` 强相关**：1080p 价格显著高于 480p / 720p；传入 `ref_videos`（带视频输入）的档位约为无视频输入的 1.5 \~ 2 倍。暂不支持音频和视频同时输出。
  </Tip>
</ParamField>

<ParamField body="prompt" type="string" required>
  文本提示词，最长 **1280 tokens**

  建议详细描述场景、主体、动作、风格等，以获得更好的生成效果。

  使用 `ref_images` / `ref_videos` / `mid_frame_images` 时，`prompt` 中**必须包含** 对应的 `@tag`（如 `@Actor-1`、`@video1`、`@image1`）。

  示例：`"@Actor-1 walks through a neon-lit street at night."`
</ParamField>

<ParamField body="duration" type="integer" default="5">
  输出视频时长（秒）

  * 取值范围：`[3, 15]`
  * 默认值：`5`

  <Warning>
    当传入 `ref_videos.type=reference` 时，`duration` 会被参考视频长度覆盖（上限 10 秒）。
  </Warning>
</ParamField>

<ParamField body="resolution" type="string" default="1080p">
  视频分辨率

  可选值：

  * `480p`
  * `720p`
  * `1080p`（默认）
</ParamField>

<ParamField body="aspect_ratio" type="string" default="16:9">
  宽高比

  可选值：

  * `16:9`（默认）
  * `4:3`
  * `1:1`
  * `9:16`
  * `3:4`

  <Warning>
    **I2V 模式下 `aspect_ratio` 会被忽略**（输出比例由输入图决定）；Omni 带 `ref_videos` 时同样忽略。
  </Warning>
</ParamField>

<ParamField body="prompt_optimizer" type="boolean" default="true">
  是否自动优化 prompt

  启用后，系统会自动优化您的提示词以获得更好的生成效果。
</ParamField>

### I2V 专用字段

<ParamField body="first_frame_image" type="string">
  视频起始帧图片 URL（jpg / jpeg / png / gif / bmp）

  传入后将以该图片作为视频的**起始画面**。
</ParamField>

<ParamField body="end_frame_image" type="string">
  视频结束帧图片 URL（jpg / jpeg / png / gif / bmp）

  传入后将以该图片作为视频的**结束画面**，可与 `first_frame_image` 组合实现首尾帧控制。
</ParamField>

<ParamField body="mid_frame_images" type="object[]">
  中间关键帧列表，最多 **6 个**。每个元素结构如下：

  <Expandable title="mid_frame_images 元素">
    <ResponseField name="tag" type="string" required>
      以 `@` 开头，且必须出现在 `prompt` 中，如 `@image1`
    </ResponseField>

    <ResponseField name="image_url" type="string" required>
      图片 URL（jpg / jpeg / png / gif / bmp）
    </ResponseField>

    <ResponseField name="time_stamp" type="integer" default="-1">
      出现时间戳（秒）。默认 `-1`（不指定）；指定时必须满足 `0 < time_stamp < duration`。
    </ResponseField>
  </Expandable>
</ParamField>

### Omni 专用字段

<ParamField body="ref_images" type="object[]">
  参考图列表（所有元素 `type` 必须一致）。每个元素结构如下：

  <Expandable title="ref_images 元素">
    <ResponseField name="tag" type="string" required>
      以 `@` 开头，且必须出现在 `prompt` 中，如 `@Actor-1`
    </ResponseField>

    <ResponseField name="type" type="string" required>
      参考类型：

      * `image` - 常规参考图（列表长度 1~~3；每项 `image_urls` 长度 1~~5）
      * `grid` - 拼贴图，即多张图合并为一张的网格图（如 4 宫格、9 宫格）；列表长度必须 = 1，`image_urls` 必须 1 张
    </ResponseField>

    <ResponseField name="image_urls" type="string[]" required>
      图片 URL 数组
    </ResponseField>

    <ResponseField name="audio_url" type="string">
      声纹音频 URL（**仅 `type=image` 支持**，音频时长 ≤ 15 秒）
    </ResponseField>
  </Expandable>
</ParamField>

<ParamField body="ref_videos" type="object[]">
  参考视频列表，**最多 1 个**。每个元素结构如下：

  <Expandable title="ref_videos 元素">
    <ResponseField name="tag" type="string" required>
      以 `@` 开头，且必须出现在 `prompt` 中，如 `@video1`
    </ResponseField>

    <ResponseField name="type" type="string" required>
      参考类型：

      * `reference` - 动作/主体参考，**会覆盖 `duration`**（跟随参考视频长度，最长 10 秒），默认带入输入视频音频；**可与 `ref_images.type=image` 组合**
      * `extend` - 视频扩展，按请求 `duration` 计费；**不能与 `ref_images` 组合**
    </ResponseField>

    <ResponseField name="video_url" type="string" required>
      视频 URL（MP4 / MOV，时长 ≤ 15 秒）
    </ResponseField>
  </Expandable>
</ParamField>

## 支持的生成场景

以下场景 `skyreels-v4-fast` 和 `skyreels-v4-std` **均支持**：

| 场景           | 模式   | 必填参数                                    | 典型用例                   |
| ------------ | ---- | --------------------------------------- | ---------------------- |
| 文生视频         | T2V  | `prompt`                                | 纯文本描述驱动，快速生成概念镜头       |
| 图生视频 - 首帧    | I2V  | `first_frame_image`                     | 静图转视频，指定起始画面           |
| 图生视频 - 尾帧    | I2V  | `end_frame_image`                       | 指定视频收束画面               |
| 图生视频 - 关键帧   | I2V  | `mid_frame_images`（1 \~ 6）              | 首 + 尾 + 中间关键帧，精准控制分镜节奏 |
| Omni 单/多主体参考 | Omni | `ref_images`（`type=image`）              | 角色一致性、多主体同框            |
| Omni 拼贴图     | Omni | `ref_images`（`type=grid`，1 张）           | 分步流程视频（教程、菜谱、操作示范）     |
| Omni 动作参考    | Omni | `ref_videos`（`type=reference`）          | 复刻参考视频的动作、主体或风格        |
| Omni 视频扩展    | Omni | `ref_videos`（`type=extend`）             | 从已有视频续写后续剧情            |
| Omni 声纹同步    | Omni | `ref_images`（`type=image`）+ `audio_url` | 数字人口播、音频驱动对口型          |

## 参数约束

以下约束违反时请求将被拒绝并返回 **422**，且 **不产生计费**：

| 参数                          | 约束                                                                     |
| --------------------------- | ---------------------------------------------------------------------- |
| `prompt`                    | 最长 1280 tokens                                                         |
| `duration`                  | `[3, 15]` 秒；带 `ref_videos.type=reference` 时被参考视频长度覆盖（上限 10 秒）          |
| `resolution`                | 仅 `480p` / `720p` / `1080p`                                            |
| `aspect_ratio`              | `16:9` / `4:3` / `1:1` / `9:16` / `3:4`；I2V 忽略；Omni 带 `ref_videos` 时忽略 |
| `mid_frame_images`          | 最多 6 个；`time_stamp` 必须为 `-1` 或落在 `(0, duration)` 区间                    |
| `ref_images` 整体             | 列表内 `type` 必须一致；不能与 I2V 字段同时出现                                         |
| `ref_images.type=grid`      | 列表长度必须 = 1；`image_urls` 必须为 1 张                                        |
| `ref_images.type=image`     | 列表长度 1 \~ 3；每项 `image_urls` 长度 1 \~ 5                                  |
| `ref_images.audio_url`      | 仅 `type=image` 支持，音频 ≤ 15 秒                                            |
| `ref_videos`                | 最多 1 个；`video_url` MP4 / MOV，≤ 15 秒                                    |
| `ref_videos.type=reference` | 覆盖请求 `duration`（最长 10 秒），可与 `ref_images.type=image` 组合，默认带入输入视频音频      |
| `ref_videos.type=extend`    | 按请求 `duration` 计费；**不能与 `ref_images` 组合**                              |
| `tag` 字段                    | 必须以 `@` 开头且出现在 `prompt` 中                                              |
| I2V / Omni 互斥               | I2V 字段与 Omni 字段不能同时出现                                                  |

## 响应

<ResponseField name="code" type="integer">
  响应状态码，成功时为 200
</ResponseField>

<ResponseField name="data" type="array">
  返回数据数组

  <Expandable title="数组元素">
    <ResponseField name="status" type="string">
      任务状态，初始提交时为 `submitted`
    </ResponseField>

    <ResponseField name="task_id" type="string">
      任务唯一标识符，用于查询任务状态和结果
    </ResponseField>
  </Expandable>
</ResponseField>

## 请求示例

### 场景 1：文生视频（最简）

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "A serene forest at sunset with golden light filtering through the trees."
}
```

### 场景 2：文生视频（完整参数）

```json theme={null}
{
  "model": "skyreels-v4-std",
  "prompt": "A serene forest at sunset.",
  "duration": 5,
  "resolution": "720p",
  "aspect_ratio": "16:9",
  "prompt_optimizer": true
}
```

### 场景 3：图生视频 - 首帧

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "Slowly pull the camera back to reveal the entire scene.",
  "first_frame_image": "https://example.com/start.png",
  "duration": 5
}
```

### 场景 4：图生视频 - 首尾帧 + 中间关键帧

```json theme={null}
{
  "model": "skyreels-v4-std",
  "prompt": "The King summons a flying dragon. @image1 The dragon lowers. The King mounts and flies away.",
  "duration": 8,
  "resolution": "1080p",
  "first_frame_image": "https://example.com/k2v_0.png",
  "end_frame_image":   "https://example.com/k2v_2.png",
  "mid_frame_images": [
    { "tag": "@image1", "image_url": "https://example.com/k2v_1.png", "time_stamp": 3 }
  ]
}
```

### 场景 5：Omni - 单主体参考

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "@Actor-1 walks through a neon-lit street at night.",
  "ref_images": [
    { "tag": "@Actor-1", "type": "image", "image_urls": ["https://example.com/actor.jpg"] }
  ]
}
```

### 场景 6：Omni - 多主体 + 视频动作参考

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "The man from @image_1 imitates the move on the left in @video_1. The woman from @image_2 imitates the right side.",
  "duration": 5,
  "ref_images": [
    { "tag": "@image_1", "type": "image", "image_urls": ["https://example.com/a.png"] },
    { "tag": "@image_2", "type": "image", "image_urls": ["https://example.com/b.png"] }
  ],
  "ref_videos": [
    { "tag": "@video_1", "type": "reference", "video_url": "https://example.com/motion.mp4" }
  ]
}
```

<Warning>
  此场景使用 `ref_videos.type=reference`，**请求的 `duration` 会被参考视频实际长度覆盖**（上限 10 秒）。即便此处传 `"duration": 5`，最终视频长度以参考视频为准。
</Warning>

### 场景 7：Omni - 拼贴图（grid）

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "Create a video showing how to make tomato and egg noodles based on @image1.",
  "ref_images": [
    { "tag": "@image1", "type": "grid", "image_urls": ["https://example.com/recipe_grid.png"] }
  ]
}
```

### 场景 8：Omni - 视频扩展（extend）

```json theme={null}
{
  "model": "skyreels-v4-fast",
  "prompt": "Video extended @video1, someone walks over and sits on the sofa.",
  "duration": 8,
  "ref_videos": [
    { "tag": "@video1", "type": "extend", "video_url": "https://example.com/source.mp4" }
  ]
}
```

### 场景 9：Omni - 带声纹（语音同步）

```json theme={null}
{
  "model": "skyreels-v4-std",
  "prompt": "@Actor-1 speaks with a calm tone.",
  "ref_images": [
    {
      "tag": "@Actor-1",
      "type": "image",
      "image_urls": ["https://example.com/actor.jpg"],
      "audio_url":  "https://example.com/voice.mp3"
    }
  ]
}
```

<Note>
  **查询任务结果**

  视频生成为异步任务，提交后会返回 `task_id`。使用 [获取任务状态](/cn/api-reference/tasks/status) 接口查询生成进度和结果。
</Note>