Simple Document Processing MCP Server

Instructions

Implementation Reference

• src/tools/pdfTools.ts:65-131 (handler)

  The core handler function that performs the PDF merging logic: reads input PDFs, copies pages using pdf-lib, saves the merged file with a unique ID in the output directory.

  export async function mergePDFs(inputPaths: string[], outputDir: string) {
    try {
      console.error(`Starting PDF merge operation...`);
      console.error(`Input files:`, inputPaths);
      console.error(`Output directory: ${outputDir}`);
  
      // 確保輸出目錄存在
      try {
        await fs.access(outputDir);
        console.error(`Output directory exists: ${outputDir}`);
      } catch {
        console.error(`Creating output directory: ${outputDir}`);
        await fs.mkdir(outputDir, { recursive: true });
        console.error(`Created output directory: ${outputDir}`);
      }
  
      const uniqueId = generateUniqueId();
      console.error(`Generated unique ID for this batch: ${uniqueId}`);
  
      // 修改輸出檔案名稱，加入 uniqueId
      const outputPath = path.join(outputDir, `merged_${uniqueId}.pdf`);
      console.error(`New output path with unique ID: ${outputPath}`);
  
      const mergedPdf = await PDFDocument.create();
  
      for (const filePath of inputPaths) {
        console.error(`Processing input file: ${filePath}`);
        const pdfBytes = await fs.readFile(filePath);
        console.error(`Read ${pdfBytes.length} bytes from ${filePath}`);
  
        const pdf = await PDFDocument.load(pdfBytes);
        const pageCount = pdf.getPageCount();
        console.error(`Loaded PDF with ${pageCount} pages from ${filePath}`);
  
        const copiedPages = await mergedPdf.copyPages(pdf, pdf.getPageIndices());
        console.error(`Copied ${copiedPages.length} pages from ${filePath}`);
  
        copiedPages.forEach((page, index) => {
          mergedPdf.addPage(page);
          console.error(`Added page ${index + 1} from ${filePath}`);
        });
      }
  
      const mergedPdfBytes = await mergedPdf.save();
      console.error(`Generated merged PDF: ${mergedPdfBytes.length} bytes`);
  
      await fs.writeFile(outputPath, mergedPdfBytes);
      console.error(`Successfully wrote merged PDF to ${outputPath}`);
  
      return {
        success: true,
        data: `Successfully merged ${inputPaths.length} PDFs into ${outputPath}`,
      };
    } catch (error) {
      console.error(`Error in mergePDFs:`);
      console.error(error);
      if (error instanceof Error) {
        console.error(`Error name: ${error.name}`);
        console.error(`Error message: ${error.message}`);
        console.error(`Error stack: ${error.stack}`);
      }
      return {
        success: false,
        error: error instanceof Error ? error.message : "Unknown error",
      };
    }
  }

• src/tools/pdfTools.ts:13-31 (schema)

  The Tool object definition for pdf_merger, including name, description, and inputSchema specifying inputPaths array and outputDir.

  export const PDF_MERGE_TOOL: Tool = {
    name: "pdf_merger",
    description: "Merge multiple PDF files into one",
    inputSchema: {
      type: "object",
      properties: {
        inputPaths: {
          type: "array",
          items: { type: "string" },
          description: "Paths to the input PDF files",
        },
        outputDir: {
          type: "string",
          description: "Directory where merged PDFs should be saved",
        },
      },
      required: ["inputPaths", "outputDir"],
    },
  };

• src/tools/_index.ts:9-9 (registration)

  Registration of pdf_merger tool (as PDF_MERGE_TOOL) in the central tools array exported for the MCP server's ListTools handler.

  export const tools = [DOCUMENT_READER_TOOL, PDF_MERGE_TOOL, PDF_SPLIT_TOOL, DOCX_TO_PDF_TOOL, DOCX_TO_HTML_TOOL, HTML_CLEAN_TOOL, HTML_TO_TEXT_TOOL, HTML_TO_MARKDOWN_TOOL, HTML_EXTRACT_RESOURCES_TOOL, HTML_FORMAT_TOOL, TEXT_DIFF_TOOL, TEXT_SPLIT_TOOL, TEXT_FORMAT_TOOL, TEXT_ENCODING_CONVERT_TOOL, EXCEL_READ_TOOL, FORMAT_CONVERTER_TOOL];

• src/index.ts:95-111 (registration)

  Dispatch/registration in the main CallToolRequestSchema handler: checks name === 'pdf_merger', extracts args, calls mergePDFs, and returns formatted response.

  if (name === "pdf_merger") {
    const { inputPaths, outputDir } = args as {
      inputPaths: string[];
      outputDir: string;
    };
    const result = await mergePDFs(inputPaths, outputDir);
    if (!result.success) {
      return {
        content: [{ type: "text", text: `Error: ${result.error}` }],
        isError: true,
      };
    }
    return {
      content: [{ type: "text", text: fileOperationResponse(result.data) }],
      isError: false,
    };
  }