MCP Simple AivisSpeech

Instructions

Implementation Reference

• src/index.ts:322-348 (handler)

  Handler for the check_engine_status tool. Uses AivisSpeechClient to check if the engine is running and returns appropriate status message.

  case 'check_engine_status': {
    try {
      const isRunning = await client.isEngineRunning();
      const version = isRunning ? await client.getVersion() : null;
  
      return {
        content: [
          {
            type: 'text',
            text: isRunning 
              ? `AivisSpeech engine is running (version: ${version})`
              : 'AivisSpeech engine is not running. Please start the engine at http://127.0.0.1:10101',
          },
        ],
      };
    } catch (error) {
      return {
        content: [
          {
            type: 'text',
            text: `Error checking engine status: ${error instanceof Error ? error.message : String(error)}`,
          },
        ],
        isError: true,
      };
    }
  }

• src/aivisspeech-client.ts:190-197 (helper)

  Supporting method isEngineRunning() in AivisSpeechClient class that checks engine status by attempting to fetch the version endpoint.

  async isEngineRunning(): Promise<boolean> {
    try {
      await this.getVersion();
      return true;
    } catch {
      return false;
    }
  }

• src/index.ts:75-82 (registration)

  Tool registration in ListTools response when engine is NOT running (only this tool is listed).

  {
    name: 'check_engine_status',
    description: 'Check if AivisSpeech engine is running',
    inputSchema: {
      type: 'object',
      properties: {},
    },
  },

• src/index.ts:161-168 (registration)

  Tool registration in ListTools response when engine IS running (listed among other tools).

  {
    name: 'check_engine_status',
    description: 'Check if AivisSpeech engine is running',
    inputSchema: {
      type: 'object',
      properties: {},
    },
  },

• src/index.ts:164-167 (schema)

  Input schema for check_engine_status tool (empty object, no parameters required).

  inputSchema: {
    type: 'object',
    properties: {},
  },