MCP Windows Website Downloader Server

Instructions

Implementation Reference

• src/mcp_windows_website_downloader/server.py:62-114 (handler)

  MCP call_tool handler specifically for the 'download' tool: validates arguments, creates download task, calls WebsiteDownloader.download(url), handles cancellation and returns result or error.
  @self.server.call_tool() async def handle_call_tool( name: str, arguments: Dict[str, Any] | None, progress_callback = None ) -> List[types.TextContent]: """Handle tool calls""" logger.info(f"Tool called: {name} with args {arguments}") try: if name != "download": raise ValueError(f"Unknown tool: {name}") if not arguments or "url" not in arguments: raise ValueError("URL is required") url = arguments["url"] # Create download task with progress tracking async def download_with_progress(): try: logger.info(f"Starting download of {url}") result = await self.downloader.download(url) logger.info("Download complete") return result except asyncio.CancelledError: logger.info("Download task cancelled") raise except Exception as e: logger.error(f"Download failed: {str(e)}") raise task = asyncio.create_task(download_with_progress()) self._tasks.add(task) try: result = await task finally: self._tasks.remove(task) return [types.TextContent( type="text", text=str(result) )] except asyncio.CancelledError: logger.info("Tool call cancelled") raise except Exception as e: logger.error(f"Tool error: {str(e)}") return [types.TextContent( type="text", text=f"Error: {str(e)}" )]
• src/mcp_windows_website_downloader/server.py:44-58 (schema)

  Tool schema definition including name 'download', description, and input schema requiring 'url' parameter.
  types.Tool( name="download", description="Download documentation website for RAG indexing", inputSchema={ "type": "object", "properties": { "url": { "type": "string", "description": "Documentation site URL" } }, "required": ["url"] } ) ]
• src/mcp_windows_website_downloader/downloader.py:88-157 (handler)

  Core implementation of the download functionality: creates site directory, analyzes crawl depth, downloads pages and assets recursively, saves processed HTML, creates RAG index JSON.
  async def download(self, url: str) -> Dict[str, Any]: """Download a documentation website""" try: # Reset state self.visited_urls.clear() self.current_domain = urlparse(url).netloc # Ensure we're using the configured output directory logger.info(f"Using output directory: {self.output_dir}") # Create site directory inside the output directory self.site_dir = self.output_dir / clean_filename(self.current_domain) logger.info(f"Creating site directory at: {self.site_dir}") self.site_dir.mkdir(exist_ok=True) logger.info(f"Starting download of {url} to {self.site_dir}") # Create clean directory structure assets_dir = self.site_dir / "assets" assets_dir.mkdir(exist_ok=True) for dir_name in ["css", "js", "images", "fonts", "other"]: (assets_dir / dir_name).mkdir(exist_ok=True) # Configure session headers = { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/91.0.4472.124', 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', 'Accept-Language': 'en-US,en;q=0.5' } async with aiohttp.ClientSession(headers=headers) as session: # Analyze site and set depth self.max_depth = await self._analyze_site_structure(session, url) logger.info(f"Using max depth of {self.max_depth} for this site") # Start download await self._process_page(session, url) # Create index index = { "url": url, "domain": self.current_domain, "pages": len(self.visited_urls), "path": str(self.site_dir), "max_depth_used": self.max_depth } index_path = self.site_dir / "rag_index.json" with open(index_path, "w") as f: json.dump(index, f, indent=2) logger.info(f"Download complete. {len(self.visited_urls)} pages saved to {self.site_dir}") return { "status": "success", "path": str(self.site_dir), "pages": len(self.visited_urls), "depth_used": self.max_depth } except asyncio.CancelledError: logger.info("Download cancelled") raise except Exception as e: logger.error(f"Download failed: {str(e)}", exc_info=True) return { "status": "error", "error": str(e) }
• src/mcp_windows_website_downloader/server.py:40-61 (registration)

  Registration of the 'download' tool via MCP Server's list_tools decorator, including schema.
  @self.server.list_tools() async def handle_list_tools() -> List[types.Tool]: logger.info("Listing tools") tools = [ types.Tool( name="download", description="Download documentation website for RAG indexing", inputSchema={ "type": "object", "properties": { "url": { "type": "string", "description": "Documentation site URL" } }, "required": ["url"] } ) ] logger.info(f"Returning {len(tools)} tools") return tools
• src/mcp_windows_website_downloader/utils.py:7-27 (helper)

  Utility function clean_filename used to sanitize filenames from URLs for safe filesystem storage.
  def clean_filename(filename: str) -> str: """ Convert URL or path to safe filename Args: filename: URL or path to clean Returns: Clean filename safe for filesystem """ # Remove query parameters and fragments filename = filename.split("?")[0].split("#")[0] # Replace unsafe characters unsafe = '<>:"\\/|?*' filename = "".join(c if c not in unsafe else "_" for c in filename) # Clean up multiple underscores filename = re.sub(r"_+", "_", filename) return filename.strip("_")