sing_over_instrumental

Add AI-generated vocals to instrumental tracks by providing lyrics and selecting a voice model, enabling users to create complete songs with synthesized singing.

Instructions

Add AI-generated vocals over an instrumental track

Input Schema

TableJSON Schema

Name	Required	Description
`instrumental_url`	Yes	URL of the instrumental audio file
`lyrics`	Yes	Lyrics to sing
`voice_id`	Yes	Voice model ID to use for singing (use get_all_voices to find IDs)
`webhook_url`	No	URL for callback upon completion

Implementation Reference

src/index.ts:1183-1203 (handler)

Handler function that executes the tool: validates required parameters (instrumental_url, lyrics, voice_id), forwards the request to the MusicGPT API endpoint '/sing_over_instrumental', and returns task status information.

private async handleSingOverInstrumental(args: any) {
  if (!args.instrumental_url || !args.lyrics || !args.voice_id) {
    throw new McpError(ErrorCode.InvalidParams, "instrumental_url, lyrics, and voice_id are required");
  }

  const response = await this.axiosInstance.post("/sing_over_instrumental", {
    instrumental_url: args.instrumental_url,
    lyrics: args.lyrics,
    voice_id: args.voice_id,
    webhook_url: args.webhook_url,
  });

  return {
    content: [
      {
        type: "text",
        text: `Singing over instrumental started!\n\n${JSON.stringify(response.data, null, 2)}\n\nUse get_conversion_by_id with the task_id to check the status.`,
      },
    ],
  };
}

src/index.ts:493-518 (schema)

Input schema definition for the tool, specifying required parameters: instrumental_url, lyrics, voice_id, and optional webhook_url.

{
  name: "sing_over_instrumental",
  description: "Add AI-generated vocals over an instrumental track",
  inputSchema: {
    type: "object" as const,
    properties: {
      instrumental_url: {
        type: "string",
        description: "URL of the instrumental audio file",
      },
      lyrics: {
        type: "string",
        description: "Lyrics to sing",
      },
      voice_id: {
        type: "string",
        description: "Voice model ID to use for singing (use get_all_voices to find IDs)",
      },
      webhook_url: {
        type: "string",
        description: "URL for callback upon completion",
      },
    },
    required: ["instrumental_url", "lyrics", "voice_id"],
  },
},

src/index.ts:709-710 (registration)
Registration in the tool dispatch switch statement: maps tool name to the handleSingOverInstrumental handler.
```
case "sing_over_instrumental":
  return await this.handleSingOverInstrumental(args);
```
src/index.ts:645-650 (registration)
Registration of the tools list handler, which includes the sing_over_instrumental tool schema via the TOOLS array.
```
this.server.setRequestHandler(
  ListToolsRequestSchema,
  async () => ({
    tools: TOOLS,
  })
);
```
src/index.ts:68-68 (helper)
Tool type constant used in get_conversion_by_id helper tool for polling status of sing_over_instrumental conversions.
```
"SING_OVER_INSTRUMENTAL",
```

MusicGPT MCP Server