MCP Server Template for Cursor IDE

Instructions

Implementation Reference

• mcp_hitchcode/server.py:108-184 (handler)

  The core handler function that fetches Railway CLI documentation from the given URL, parses the HTML using BeautifulSoup to extract CLI commands from headings and lists/code blocks, and returns optimized text content.

  async def fetch_railway_docs_optimized(
      url: str = "https://docs.railway.app/guides/cli",
  ) -> list[types.TextContent]:
      headers = {
          "User-Agent": "MCP Railway Docs Fetcher (github.com/modelcontextprotocol/python-sdk)"
      }
  
      try:
          timeout = httpx.Timeout(10.0, connect=5.0)
          async with httpx.AsyncClient(
              follow_redirects=True, headers=headers, timeout=timeout
          ) as client:
              response = await client.get(url)
              response.raise_for_status()
  
              # Parse HTML content
              soup = BeautifulSoup(response.text, "html.parser")
              content = []
  
              # Try to find any command sections
              for heading in soup.find_all(["h1", "h2", "h3", "h4"]):
                  heading_text = heading.get_text(strip=True).lower()
                  if any(
                      keyword in heading_text for keyword in ["command", "cli", "usage"]
                  ):
                      # Look for the next element that might contain commands
                      next_elem = heading.find_next_sibling()
                      while next_elem and next_elem.name not in [
                          "ul",
                          "ol",
                          "h1",
                          "h2",
                          "h3",
                          "h4",
                      ]:
                          next_elem = next_elem.find_next_sibling()
  
                      if next_elem and next_elem.name in ["ul", "ol"]:
                          commands = []
                          for li in next_elem.find_all("li"):
                              cmd_text = li.get_text(strip=True)
                              if cmd_text:  # Only add non-empty commands
                                  commands.append(cmd_text)
  
                          if commands:  # Only add sections that have commands
                              content.append(f"\n{heading.get_text(strip=True)}:")
                              content.extend(f"- {cmd}" for cmd in commands)
  
              if content:
                  return [types.TextContent(type="text", text="\n".join(content))]
  
              # If we couldn't find any command sections, try to extract any code blocks
              code_blocks = soup.find_all(["pre", "code"])
              if code_blocks:
                  content = ["Railway CLI Commands:"]
                  for block in code_blocks:
                      code_text = block.get_text(strip=True)
                      if code_text and any(
                          keyword in code_text.lower() for keyword in ["railway", "cli"]
                      ):
                          content.append(code_text)
                  return [types.TextContent(type="text", text="\n".join(content))]
  
              # If still nothing found, return a more helpful message
              return [
                  types.TextContent(
                      type="text",
                      text="Could not find any CLI commands in the documentation. The page structure might have changed.",
                  )
              ]
      except Exception as e:
          return [
              types.TextContent(
                  type="text",
                  text=f"Error: Failed to fetch or parse Railway CLI docs: {str(e)}",
              )
          ]

• mcp_hitchcode/server.py:436-438 (registration)

  Dispatch registration in the @app.call_tool() handler (fetch_tool function) that routes calls to the fetch_railway_docs_optimized function when the tool name matches.

  if name == "fetch_railway_docs_optimized":
      url = arguments.get("url", "https://docs.railway.app/guides/cli")
      return await fetch_railway_docs_optimized(url)

• mcp_hitchcode/server.py:602-614 (schema)

  Tool schema and metadata registration in the @app.list_tools() function, defining the tool name, description, and optional 'url' input schema.

  types.Tool(
      name="fetch_railway_docs_optimized",
      description="Fetches the most recent Railway CLI documentation. Optionally, provide a custom URL.",
      inputSchema={
          "type": "object",
          "properties": {
              "url": {
                  "type": "string",
                  "description": "Optional custom URL for fetching Railway CLI docs.",
              },
          },
      },
  ),