MCP2Brave

Instructions

Implementation Reference

• mcp2brave.py:441-450 (handler)

  The handler function for the 'url_content' MCP tool. It takes a URL as input and returns the webpage content by calling the internal helper _get_url_content_direct.
  def url_content(url: str) -> str: """直接获取网页内容 参数: url (str): 目标网页地址 返回: str: 网页内容和元数据 """ return _get_url_content_direct(url)
• mcp2brave.py:440-440 (registration)

  The @mcp.tool() decorator registers the url_content function as an MCP tool in the FastMCP server.
  @mcp.tool()
• mcp2brave.py:282-355 (helper)

  Core helper function that implements the logic to fetch the URL content using requests, parse HTML with BeautifulSoup, extract main text content, clean it, and return with metadata. This is called by the url_content handler.
  def _get_url_content_direct(url: str) -> str: """Internal function to get content directly using requests""" try: logger.debug(f"Directly fetching content from URL: {url}") response = requests.get(url, timeout=10, headers={ 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36' }) response.raise_for_status() # 尝试检测编码 if 'charset' in response.headers.get('content-type', '').lower(): response.encoding = response.apparent_encoding try: from bs4 import BeautifulSoup soup = BeautifulSoup(response.text, 'html.parser') # 移除不需要的元素 for element in soup(['script', 'style', 'header', 'footer', 'nav', 'aside', 'iframe', 'ad', '.advertisement']): element.decompose() # 尝试找到主要内容区域 main_content = None possible_content_elements = [ soup.find('article'), soup.find('main'), soup.find(class_='content'), soup.find(id='content'), soup.find(class_='post-content'), soup.find(class_='article-content'), soup.find(class_='entry-content'), soup.find(class_='main-content'), soup.select_one('div[class*="content"]'), # 包含 "content" 的任何 class ] for element in possible_content_elements: if element: main_content = element break if not main_content: main_content = soup text = main_content.get_text(separator='\n') lines = [] for line in text.split('\n'): line = line.strip() if line and len(line) > 30: lines.append(line) cleaned_text = ' '.join(lines) if len(cleaned_text) > 1000: end_pos = cleaned_text.rfind('. ', 0, 1000) if end_pos > 0: cleaned_text = cleaned_text[:end_pos + 1] else: cleaned_text = cleaned_text[:1000] metadata = f"URL: {url}\n" metadata += f"Content Length: {len(response.text)} characters\n" metadata += f"Content Type: {response.headers.get('content-type', 'Unknown')}\n" metadata += "---\n\n" return f"{metadata}{cleaned_text}" except Exception as e: logger.error(f"Error extracting text from HTML: {str(e)}") return f"Error extracting text: {str(e)}" except Exception as e: logger.error(f"Error fetching URL content directly: {str(e)}") return f"Error getting content: {str(e)}"