AWS Documentation MCP Server

Instructions

Implementation Reference

• awslabs/aws_documentation_mcp_server/server_aws_cn.py:66-135 (handler)

  Handler for read_documentation tool in AWS China server, includes registration via @mcp.tool(), input schema via Pydantic Fields, validation, and delegation to impl.
  @mcp.tool() async def read_documentation( ctx: Context, url: Union[AnyUrl, str] = Field(description='URL of the AWS China documentation page to read'), max_length: int = Field( default=5000, description='Maximum number of characters to return.', gt=0, lt=1000000, ), start_index: int = Field( default=0, description='On return output starting at this character index, useful if a previous fetch was truncated and more content is required.', ge=0, ), ) -> str: """Fetch and convert an AWS China documentation page to markdown format. ## Usage This tool retrieves the content of an AWS China documentation page and converts it to markdown format. For long documents, you can make multiple calls with different start_index values to retrieve the entire content in chunks. ## URL Requirements - Must be from the docs.amazonaws.cn domain - Must end with .html ## Example URLs - https://docs.amazonaws.cn/en_us/AmazonS3/latest/userguide/bucketnamingrules.html - https://docs.amazonaws.cn/en_us/lambda/latest/dg/lambda-invocation.html ## Output Format The output is formatted as markdown text with: - Preserved headings and structure - Code blocks for examples - Lists and tables converted to markdown format ## Handling Long Documents If the response indicates the document was truncated, you have several options: 1. **Continue Reading**: Make another call with start_index set to the end of the previous response 2. **Stop Early**: For very long documents (>30,000 characters), if you've already found the specific information needed, you can stop reading Args: ctx: MCP context for logging and error handling url: URL of the AWS China documentation page to read max_length: Maximum number of characters to return start_index: On return output starting at this character index Returns: Markdown content of the AWS China documentation """ # Validate that URL is from docs.amazonaws.cn and ends with .html url_str = str(url) if not re.match(r'^https?://docs\.amazonaws\.cn/', url_str): error_msg = f'Invalid URL: {url_str}. URL must be from the docs.amazonaws.cn domain' await ctx.error(error_msg) return error_msg if not url_str.endswith('.html'): error_msg = f'Invalid URL: {url_str}. URL must end with .html' await ctx.error(error_msg) return error_msg return await read_documentation_impl(ctx, url_str, max_length, start_index, SESSION_UUID)
• awslabs/aws_documentation_mcp_server/server_aws.py:77-144 (handler)

  Handler for read_documentation tool in global AWS server, includes registration via @mcp.tool(), input schema via Pydantic Fields, validation, and delegation to impl.
  @mcp.tool() async def read_documentation( ctx: Context, url: str = Field(description='URL of the AWS documentation page to read'), max_length: int = Field( default=5000, description='Maximum number of characters to return.', gt=0, lt=1000000, ), start_index: int = Field( default=0, description='On return output starting at this character index, useful if a previous fetch was truncated and more content is required.', ge=0, ), ) -> str: """Fetch and convert an AWS documentation page to markdown format. ## Usage This tool retrieves the content of an AWS documentation page and converts it to markdown format. For long documents, you can make multiple calls with different start_index values to retrieve the entire content in chunks. ## URL Requirements - Must be from the docs.aws.amazon.com domain - Must end with .html ## Example URLs - https://docs.aws.amazon.com/AmazonS3/latest/userguide/bucketnamingrules.html - https://docs.aws.amazon.com/lambda/latest/dg/lambda-invocation.html ## Output Format The output is formatted as markdown text with: - Preserved headings and structure - Code blocks for examples - Lists and tables converted to markdown format ## Handling Long Documents If the response indicates the document was truncated, you have several options: 1. **Continue Reading**: Make another call with start_index set to the end of the previous response 2. **Stop Early**: For very long documents (>30,000 characters), if you've already found the specific information needed, you can stop reading Args: ctx: MCP context for logging and error handling url: URL of the AWS documentation page to read max_length: Maximum number of characters to return start_index: On return output starting at this character index Returns: Markdown content of the AWS documentation """ # Validate that URL is from docs.aws.amazon.com and ends with .html url_str = str(url) if not re.match(r'^https?://docs\.aws\.amazon\.com/', url_str): await ctx.error(f'Invalid URL: {url_str}. URL must be from the docs.aws.amazon.com domain') raise ValueError('URL must be from the docs.aws.amazon.com domain') if not url_str.endswith('.html'): await ctx.error(f'Invalid URL: {url_str}. URL must end with .html') raise ValueError('URL must end with .html') return await read_documentation_impl(ctx, url_str, max_length, start_index, SESSION_UUID)
• awslabs/aws_documentation_mcp_server/server_utils.py:33-85 (helper)

  Shared helper function containing the core logic for fetching, extracting, and formatting documentation content.
  async def read_documentation_impl( ctx: Context, url_str: str, max_length: int, start_index: int, session_uuid: str, ) -> str: """The implementation of the read_documentation tool.""" logger.debug(f'Fetching documentation from {url_str}') url_with_session = f'{url_str}?session={session_uuid}' async with httpx.AsyncClient() as client: try: response = await client.get( url_with_session, follow_redirects=True, headers={ 'User-Agent': DEFAULT_USER_AGENT, 'X-MCP-Session-Id': session_uuid, }, timeout=30, ) except httpx.HTTPError as e: error_msg = f'Failed to fetch {url_str}: {str(e)}' logger.error(error_msg) await ctx.error(error_msg) return error_msg if response.status_code >= 400: error_msg = f'Failed to fetch {url_str} - status code {response.status_code}' logger.error(error_msg) await ctx.error(error_msg) return error_msg page_raw = response.text content_type = response.headers.get('content-type', '') if is_html_content(page_raw, content_type): content = extract_content_from_html(page_raw) else: content = page_raw result = format_documentation_result(url_str, content, start_index, max_length) # Log if content was truncated if len(content) > start_index + max_length: logger.debug( f'Content truncated at {start_index + max_length} of {len(content)} characters' ) return result