Office Word MCP Server

Instructions

Implementation Reference

• word_document_server/tools/comment_tools.py:20-56 (handler)

  Primary handler function for the get_all_comments tool. Loads the Word document, extracts all comments using the core helper, handles errors, and returns a JSON-formatted response.

  async def get_all_comments(filename: str) -> str:
      """
      Extract all comments from a Word document.
      
      Args:
          filename: Path to the Word document
          
      Returns:
          JSON string containing all comments with metadata
      """
      filename = ensure_docx_extension(filename)
      
      if not os.path.exists(filename):
          return json.dumps({
              'success': False,
              'error': f'Document {filename} does not exist'
          }, indent=2)
      
      try:
          # Load the document
          doc = Document(filename)
          
          # Extract all comments
          comments = extract_all_comments(doc)
          
          # Return results
          return json.dumps({
              'success': True,
              'comments': comments,
              'total_comments': len(comments)
          }, indent=2)
          
      except Exception as e:
          return json.dumps({
              'success': False,
              'error': f'Failed to extract comments: {str(e)}'
          }, indent=2)

• word_document_server/main.py:403-406 (registration)

  MCP tool registration using @mcp.tool() decorator. This is the entry point handler that delegates to the implementation in comment_tools.

  @mcp.tool()
  def get_all_comments(filename: str):
      """Extract all comments from a Word document."""
      return comment_tools.get_all_comments(filename)

• word_document_server/core/comments.py:14-56 (helper)

  Core helper function that performs the low-level extraction of all comments from the docx Document object, handling both direct comments part and fallback methods.

  def extract_all_comments(doc: DocumentType) -> List[Dict[str, Any]]:
      """
      Extract all comments from a Word document.
      
      Args:
          doc: The Document object to extract comments from
          
      Returns:
          List of dictionaries containing comment information
      """
      comments = []
      
      # Access the document's comment part if it exists
      try:
          # Get the document part
          document_part = doc.part
          
          # Find comments part through relationships
          comments_part = None
          for rel_id, rel in document_part.rels.items():
              if 'comments' in rel.reltype and 'comments' == rel.reltype.split('/')[-1]:
                  comments_part = rel.target_part
                  break
          
          if comments_part:
              # Extract comments from the comments part using proper xpath syntax
              comment_elements = comments_part.element.xpath('.//w:comment')
              
              for idx, comment_element in enumerate(comment_elements):
                  comment_data = extract_comment_data(comment_element, idx)
                  if comment_data:
                      comments.append(comment_data)
          
          # If no comments found, try alternative approach
          if not comments:
              # Fallback: scan paragraphs for comment references
              comments = extract_comments_from_paragraphs(doc)
      
      except Exception as e:
          # If direct access fails, try alternative approach
          comments = extract_comments_from_paragraphs(doc)
      
      return comments

• word_document_server/tools/__init__.py:40-42 (helper)

  Export of the get_all_comments function from comment_tools module, making it available for import.

  from word_document_server.tools.comment_tools import (
      get_all_comments, get_comments_by_author, get_comments_for_paragraph
  )