PPT-MCP

Instructions

Implementation Reference

• src/tools/ppt-reader.ts:37-86 (handler)

  The async run() method implementing the core tool execution logic, which validates the file path, checks existence, and returns file metadata with a note on reading limitations due to PptxGenJS.

  async run(args: { file_path: string; output_format?: string; slide_range?: { start?: number; end?: number }; include_notes?: boolean }) {
    try {
      // Parameter validation
      if (!args.file_path) {
        throw new Error("File path is required");
      }
      
      if (!fs.existsSync(args.file_path)) {
        throw new Error(`File not found: ${args.file_path}`);
      }
      
      // Note: PptxGenJS doesn't support reading existing PowerPoint files
      // This is a limitation of the library - it's primarily for creation, not reading
      // For reading PPT files, we'd need a different library like 'officegen' or 'node-pptx'
      
      const fileStats = fs.statSync(args.file_path);
      const fileName = path.basename(args.file_path);
      const fileSize = (fileStats.size / 1024).toFixed(2); // KB
      
      return {
        content: [{
          type: "text",
          text: `⚠️ **Read Operation Not Fully Supported**\n\n` +
                `Unfortunately, PptxGenJS doesn't support reading existing PowerPoint files. ` +
                `This is a current limitation of the library.\n\n` +
                `**File Information:**\n` +
                `📄 **Name:** ${fileName}\n` +
                `📁 **Path:** ${args.file_path}\n` +
                `📊 **Size:** ${fileSize} KB\n` +
                `📅 **Modified:** ${fileStats.mtime.toLocaleString()}\n\n` +
                `**Alternative approaches:**\n` +
                `• Use Microsoft PowerPoint to view the content\n` +
                `• Convert to PDF and use PDF reading tools\n` +
                `• Use online PowerPoint viewers\n` +
                `• Consider using a different library like 'mammoth' for basic text extraction\n\n` +
                `**Requested format:** ${args.output_format || 'text'}\n` +
                `**Include notes:** ${args.include_notes ? 'Yes' : 'No'}`
        }]
      };
      
    } catch (error) {
      return {
        content: [{
          type: "text",
          text: `❌ **Failed to read presentation:** ${error instanceof Error ? error.message : String(error)}`
        }],
        isError: true
      };
    }
  }

• src/tools/ppt-reader.ts:7-35 (schema)

  JSON schema defining the input parameters for the read_presentation tool, including file_path (required), output_format, slide_range, and include_notes.

  parameters: {
    type: "object",
    properties: {
      file_path: {
        type: "string",
        description: "Path to the PowerPoint file to read"
      },
      output_format: {
        type: "string",
        description: "Output format for the content",
        enum: ["text", "json", "markdown"],
        default: "text"
      },
      slide_range: {
        type: "object",
        description: "Range of slides to read (optional)",
        properties: {
          start: { type: "number", description: "Start slide number (1-based)" },
          end: { type: "number", description: "End slide number (1-based)" }
        }
      },
      include_notes: {
        type: "boolean",
        description: "Whether to include speaker notes",
        default: false
      }
    },
    required: ["file_path"]
  },

• src/index.ts:32-36 (registration)

  Registration of the read_presentation tool in the ListToolsRequestSchema handler, exposing name, description, and input schema.

  {
    name: pptReader.name,
    description: pptReader.description,
    inputSchema: pptReader.parameters
  },

• src/index.ts:63-64 (registration)

  Registration in the CallToolRequestSchema switch statement, dispatching tool calls to pptReader.run().

  case "read_presentation":
    return await pptReader.run(request.params.arguments as any || {});