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

# 人声分离与分轨

> 工具箱模式填写音频 URL；原生分轨选择本站歌曲 ID，按返回轨道分别试听和下载。

> 能力按服务端配置逐项启用。未启用操作返回 503 且不创建任务；音乐积分扣除当前未启用。

## 调用步骤

工具箱模式填写音频 URL；原生分轨选择本站歌曲 ID，按返回轨道分别试听和下载。

* 1. 按下方示例准备输入，将占位地址和 ID 换成你有权使用的资源。
* 2. 将 model 设为 suno\_separate\_vocals，参数放在 input 内。提交一次并保存返回的 task\_id。
* 3. 查询已保存的任务，或接收回调。任务仍在处理中时继续等待，不重新提交同一操作。
* 遍历 result.stems，读取各轨的 download\_url 和到期时间，以实际返回的轨名和数量为准。

## 创建任务

将model和input发送到统一任务入口；可在顶层传公开HTTPS callback\_url。资源ID均为本账户的本站ID。当前音乐积分扣除未启用。

```json theme={null}
{
  "model": "suno_separate_vocals",
  "input": {
    "type": "separate_vocal",
    "upload_url": "https://example.com/song.mp3"
  }
}
```

## 受理响应

HTTP 200仅表示受理。查询GET /api/v1/tasks/\{task\_id}，直到data.status为succeeded或failed。成功后读取data.result；失败原因位于data.message。

```json theme={null}
{
  "code": 200,
  "msg": "success",
  "data": {
    "task_id": "YOUR_TASK_ID"
  }
}
```

## 参数：type = separate\_vocal

| 字段 | 类型 | 必填 | 说明与限制 |
| - | - | - | - |
| type | string | 是 | separate\_vocal（2 轨）、split\_stem（4/6 轨），或原生 split\_stem\_advanced。 = separate\_vocal |
| upload\_url | string | 是 | 公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。 |
| output\_format | string | 否 / 依条件 | 工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。 values: mp3, wav, flac, ogg, m4a, aac; default: wav |

## 参数：type = split\_stem

| 字段 | 类型 | 必填 | 说明与限制 |
| - | - | - | - |
| type | string | 是 | separate\_vocal（2 轨）、split\_stem（4/6 轨），或原生 split\_stem\_advanced。 = split\_stem |
| upload\_url | string | 是 | 公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。 |
| stem\_mode | number | 是 | 工具箱分轨：4 或 6 轨；人声分离在后台固定为 2 轨。 values: 4, 6 |
| output\_format | string | 否 / 依条件 | 工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。 values: mp3, wav, flac, ogg, m4a, aac; default: wav |

## 参数：type = split\_stem\_advanced

| 字段 | 类型 | 必填 | 说明与限制 |
| - | - | - | - |
| type | string | 是 | separate\_vocal（2 轨）、split\_stem（4/6 轨），或原生 split\_stem\_advanced。 = split\_stem\_advanced |
| audio\_id | string | 是 | 本账户的本站歌曲ID。翻唱/续写只传此ID，不传源task\_id或外部URL。 maxLength: 200 |
| stem\_name | string | 否 / 依条件 | 原生抽取的轨道名。当前仅 Piano 有明确文档并开放；省略时使用十二轨分离。 maxLength: 100 |

## 输入与限制

下方Schema列出支持字段、类型和基础限制；还须遵守“重要条件”，跨字段与来源资源校验不完全体现在JSON Schema中。未知参数会被拒绝。URL 参数仅接受公开 HTTPS 地址，不接受文件对象。

```json theme={null}
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "oneOf": [
    {
      "type": "object",
      "properties": {
        "type": {
          "type": "string",
          "const": "separate_vocal",
          "description": "separate_vocal（2 轨）、split_stem（4/6 轨），或原生 split_stem_advanced。"
        },
        "upload_url": {
          "type": "string",
          "format": "uri",
          "description": "公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。"
        },
        "output_format": {
          "default": "wav",
          "type": "string",
          "enum": [
            "mp3",
            "wav",
            "flac",
            "ogg",
            "m4a",
            "aac"
          ],
          "description": "工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。"
        }
      },
      "required": [
        "type",
        "upload_url"
      ],
      "additionalProperties": false
    },
    {
      "type": "object",
      "properties": {
        "type": {
          "type": "string",
          "const": "split_stem",
          "description": "separate_vocal（2 轨）、split_stem（4/6 轨），或原生 split_stem_advanced。"
        },
        "upload_url": {
          "type": "string",
          "format": "uri",
          "description": "公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。"
        },
        "stem_mode": {
          "anyOf": [
            {
              "type": "number",
              "const": 4
            },
            {
              "type": "number",
              "const": 6
            }
          ],
          "description": "工具箱分轨：4 或 6 轨；人声分离在后台固定为 2 轨。"
        },
        "output_format": {
          "default": "wav",
          "type": "string",
          "enum": [
            "mp3",
            "wav",
            "flac",
            "ogg",
            "m4a",
            "aac"
          ],
          "description": "工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。"
        }
      },
      "required": [
        "type",
        "upload_url",
        "stem_mode"
      ],
      "additionalProperties": false
    },
    {
      "type": "object",
      "properties": {
        "type": {
          "type": "string",
          "const": "split_stem_advanced",
          "description": "separate_vocal（2 轨）、split_stem（4/6 轨），或原生 split_stem_advanced。"
        },
        "audio_id": {
          "type": "string",
          "minLength": 1,
          "maxLength": 200,
          "description": "本账户的本站歌曲ID。翻唱/续写只传此ID，不传源task_id或外部URL。"
        },
        "stem_name": {
          "type": "string",
          "minLength": 1,
          "maxLength": 100,
          "description": "原生抽取的轨道名。当前仅 Piano 有明确文档并开放；省略时使用十二轨分离。"
        }
      },
      "required": [
        "type",
        "audio_id"
      ],
      "additionalProperties": false
    }
  ]
}
```

## 重要条件

* separate\_vocal与split\_stem使用upload\_url，音频上限100 MiB。separate\_vocal请求两轨；split\_stem必须传stem\_mode=4或6。output\_format使用所列格式，默认wav。
* 原生split\_stem\_advanced传audio\_id，不传upload\_url或output\_format；省略stem\_name为十二轨分离，单轨抽取当前仅开放Piano，其他名称返回42201。结果可能包含多组轨道，请遍历实际stems数组，不假定固定条数。
* 音频URL必须包含可读且时长大于零的MP3/WAV/FLAC/OGG/M4A/AAC音轨，每个文件最多100 MiB。URL须为可公开读取的HTTPS地址、端口443，不含用户名密码或私网地址，且应持续可访问直到任务结束。
* 上传文件保留24小时，生成文件保留14天，任务详情保留30天；复用文件沿用原到期时间。有效Voice/Persona身份映射与任务日志清理分开处理，请及时下载需长期保存的文件。

## 结果 Schema

创建响应为 \{code:200,msg:"success",data:\{task\_id}}，轮询 GET /api/v1/tasks/\{task\_id}。成功时 data.result 为符合下方 Schema 的对象，其他状态为 null；input 仍为 JSON 字符串。

```json theme={null}
{
  "type": "object",
  "required": [
    "stems"
  ],
  "properties": {
    "stems": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "stem_name": {
            "type": [
              "string",
              "null"
            ]
          },
          "download_url": {
            "type": "string"
          },
          "expires_at": {
            "type": "string"
          }
        }
      }
    }
  },
  "additionalProperties": false
}
```

## 相关流程

[Voice 完整四步](https://docs.apistable.com/zh/suno-voice-workflow.md)
[文件上传](https://docs.apistable.com/zh/file-upload.md)
[回调](https://docs.apistable.com/zh/webhooks.md)


## OpenAPI

````yaml openapi-suno-separate-vocals-zh.json POST /api/v1/tasks/create_task
openapi: 3.1.0
info:
  title: APIstable API
  version: 1.0.0-preview
  description: >-
    Public /api/v1 protocol. Valid API keys have the same account-scoped access.
    No client submission idempotency or max_credits. Enabled Suno operations use
    PayAPI and durable task processing. Media uses R2: uploads 24 hours,
    generated files 14 days, task logs 30 days. Music credit deduction is
    disabled. Unconfigured operations and LLM return 503.
servers:
  - url: https://apistable.com
    description: APIstable API
security: []
paths:
  /api/v1/tasks/create_task:
    post:
      summary: 人声分离与分轨
      description: 工具箱模式填写音频 URL；原生分轨选择本站歌曲 ID，按返回轨道分别试听和下载。
      operationId: suno_separate_vocals
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
                - model
                - input
              additionalProperties: false
              properties:
                model:
                  type: string
                  enum:
                    - suno_separate_vocals
                  example: suno_separate_vocals
                input:
                  $schema: https://json-schema.org/draft/2020-12/schema
                  oneOf:
                    - type: object
                      properties:
                        type:
                          type: string
                          const: separate_vocal
                          description: >-
                            separate_vocal（2 轨）、split_stem（4/6 轨），或原生
                            split_stem_advanced。
                        upload_url:
                          type: string
                          format: uri
                          description: >-
                            公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。
                        output_format:
                          default: wav
                          type: string
                          enum:
                            - mp3
                            - wav
                            - flac
                            - ogg
                            - m4a
                            - aac
                          description: 工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。
                      required:
                        - type
                        - upload_url
                      additionalProperties: false
                    - type: object
                      properties:
                        type:
                          type: string
                          const: split_stem
                          description: >-
                            separate_vocal（2 轨）、split_stem（4/6 轨），或原生
                            split_stem_advanced。
                        upload_url:
                          type: string
                          format: uri
                          description: >-
                            公开HTTPS音频URL，端口443且不含用户名密码。接受包含可读音轨的MP3/WAV/FLAC/OGG/M4A/AAC，大小限制见具体操作。
                        stem_mode:
                          anyOf:
                            - type: number
                              const: 4
                            - type: number
                              const: 6
                          description: 工具箱分轨：4 或 6 轨；人声分离在后台固定为 2 轨。
                        output_format:
                          default: wav
                          type: string
                          enum:
                            - mp3
                            - wav
                            - flac
                            - ogg
                            - m4a
                            - aac
                          description: 工具箱输出：mp3、wav（默认）、flac、ogg、m4a 或 aac；原生轨道抽取不接受此参数。
                      required:
                        - type
                        - upload_url
                        - stem_mode
                      additionalProperties: false
                    - type: object
                      properties:
                        type:
                          type: string
                          const: split_stem_advanced
                          description: >-
                            separate_vocal（2 轨）、split_stem（4/6 轨），或原生
                            split_stem_advanced。
                        audio_id:
                          type: string
                          minLength: 1
                          maxLength: 200
                          description: 本账户的本站歌曲ID。翻唱/续写只传此ID，不传源task_id或外部URL。
                        stem_name:
                          type: string
                          minLength: 1
                          maxLength: 100
                          description: 原生抽取的轨道名。当前仅 Piano 有明确文档并开放；省略时使用十二轨分离。
                      required:
                        - type
                        - audio_id
                      additionalProperties: false
                callback_url:
                  type: string
                  format: uri
                  description: >-
                    Optional public HTTPS endpoint for signed completion
                    callbacks; no redirects or private IPs.
            example:
              model: suno_separate_vocals
              input:
                type: separate_vocal
                upload_url: https://example.com/song.mp3
      responses:
        '200':
          description: Accepted task. Poll the common task endpoint or use callback_url.
          content:
            application/json:
              schema:
                type: object
                required:
                  - code
                  - msg
                  - data
                additionalProperties: false
                properties:
                  code:
                    const: 200
                  msg:
                    const: success
                  data:
                    type: object
                    required:
                      - task_id
                    additionalProperties: false
                    properties:
                      task_id:
                        type: string
        '400':
          description: >-
            40001 invalid JSON, 40002 invalid input/callback, 40003 unsupported
            parameter.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '401':
          description: Invalid or revoked API key.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '402':
          description: '40201: Insufficient credits (live acceptance).'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '404':
          description: '40401: Unknown model or resource not found in this account.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '409':
          description: '40901: Resource belongs to a different provider account.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '410':
          description: '41001: Source audio or temporary resource has expired.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '413':
          description: '41301: Request too large.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '422':
          description: '42201: Requested music combination is not enabled.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '429':
          description: '42901: Rate/concurrency limit (live acceptance).'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
        '503':
          description: '50301: Service/provider unavailable. No live submission or charge.'
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Error'
      security:
        - bearerAuth: []
components:
  schemas:
    Error:
      type: object
      required:
        - code
        - msg
        - data
      properties:
        code:
          type: integer
        msg:
          type: string
        data:
          type: 'null'
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: >-
        Every valid key can submit and query its own account. Revoked keys are
        rejected.

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.