Advanced TTS MCP Server

Instructions

Implementation Reference

• src/typescript/index.ts:499-535 (handler)

  The handler function in AdvancedTTSServer class that executes the tool logic: reads the output directory, filters for supported audio formats (wav, mp3, flac, ogg), retrieves file stats, and returns sorted list with metadata.

  async listOutputFiles() {
    try {
      const files = await fs.readdir(this.outputDir);
      const audioFiles = files.filter(file => 
        AudioFormats.some(format => file.endsWith(`.${format}`))
      );
  
      const fileInfo = await Promise.all(
        audioFiles.map(async (file) => {
          const filePath = join(this.outputDir, file);
          const stats = await fs.stat(filePath);
          return {
            filename: file,
            path: filePath,
            size: stats.size,
            created: stats.birthtime.toISOString(),
            modified: stats.mtime.toISOString(),
            format: file.split('.').pop()
          };
        })
      );
  
      return {
        success: true,
        files: fileInfo.sort((a, b) => 
          new Date(b.modified).getTime() - new Date(a.modified).getTime()
        ),
        totalFiles: fileInfo.length,
        outputDirectory: this.outputDir
      };
    } catch (error) {
      return {
        success: false,
        error: `Failed to list files: ${error}`
      };
    }
  }

• src/typescript/index.ts:722-728 (registration)

  Registers the 'list_output_files' tool in the server's listTools response, including name, description, and input schema (no required parameters).

    name: 'list_output_files',
    description: 'List saved audio files in the output directory with metadata',
    inputSchema: {
      type: 'object',
      properties: {},
    },
  },

• src/typescript/index.ts:725-727 (schema)

  Defines the input schema for the tool: an empty object since no parameters are required.

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

• src/typescript/index.ts:820-846 (helper)

  MCP server request handler that calls the listOutputFiles method and formats the response content for the tool call, including error handling and rich text formatting.

  case 'list_output_files':
    const fileList = await ttsServer.listOutputFiles();
    if (fileList.success) {
      const files = fileList.files?.map(f => 
        `**${f.filename}** (${(f.size / 1024).toFixed(1)}KB) - ${f.format?.toUpperCase()}`
      ).join('\n');
      
      return {
        content: [
          {
            type: 'text',
            text: `🎙️ **Output Files (${fileList.totalFiles})**\n\n` +
                 `**Directory:** ${fileList.outputDirectory}\n\n` +
                 `${files || 'No audio files found'}`,
          },
        ],
      };
    } else {
      return {
        content: [
          {
            type: 'text',
            text: `❌ **Error:** ${fileList.error}`,
          },
        ],
      };
    }