mcp-server-fetch 0.1.0__py3-none-any.whl → 0.6.1__py3-none-any.whl

mcp_server_fetch/__init__.py

@@ -3,9 +3,21 @@ from .server import serve
 
 def main():
     """MCP Fetch Server - HTTP fetching functionality for MCP"""
+    import argparse
     import asyncio
 
-    asyncio.run(serve())
+    parser = argparse.ArgumentParser(
+        description="give a model the ability to make web requests"
+    )
+    parser.add_argument("--user-agent", type=str, help="Custom User-Agent string")
+    parser.add_argument(
+        "--ignore-robots-txt",
+        action="store_true",
+        help="Ignore robots.txt restrictions",
+    )
+
+    args = parser.parse_args()
+    asyncio.run(serve(args.user_agent, args.ignore_robots_txt))
 
 
 if __name__ == "__main__":

mcp_server_fetch/server.py

@@ -1,53 +1,199 @@
+from typing import Annotated, Tuple
+from urllib.parse import urlparse, urlunparse
+
 import markdownify
 import readabilipy.simple_json
+from mcp.shared.exceptions import McpError
 from mcp.server import Server
 from mcp.server.stdio import stdio_server
 from mcp.types import (
-    TextContent,
-    Tool,
+    GetPromptResult,
     Prompt,
     PromptArgument,
-    GetPromptResult,
     PromptMessage,
+    TextContent,
+    Tool,
+    INVALID_PARAMS,
+    INTERNAL_ERROR,
 )
-from pydantic import BaseModel, Field
+from protego import Protego
+from pydantic import BaseModel, Field, AnyUrl
+
+DEFAULT_USER_AGENT_AUTONOMOUS = "ModelContextProtocol/1.0 (Autonomous; +https://github.com/modelcontextprotocol/servers)"
+DEFAULT_USER_AGENT_MANUAL = "ModelContextProtocol/1.0 (User-Specified; +https://github.com/modelcontextprotocol/servers)"
+
 
+def extract_content_from_html(html: str) -> str:
+    """Extract and convert HTML content to Markdown format.
 
-def extract_content(html: str) -> str:
-    ret = readabilipy.simple_json.simple_json_from_html_string(html)
-    if not ret["plain_content"]:
+    Args:
+        html: Raw HTML content to process
+
+    Returns:
+        Simplified markdown version of the content
+    """
+    ret = readabilipy.simple_json.simple_json_from_html_string(
+        html, use_readability=True
+    )
+    if not ret["content"]:
         return "<error>Page failed to be simplified from HTML</error>"
     content = markdownify.markdownify(
-        ret["plain_content"],
+        ret["content"],
         heading_style=markdownify.ATX,
     )
     return content
 
 
-async def fetch_url(url: str) -> str:
-    from httpx import AsyncClient
+def get_robots_txt_url(url: AnyUrl | str) -> str:
+    """Get the robots.txt URL for a given website URL.
+
+    Args:
+        url: Website URL to get robots.txt for
+
+    Returns:
+        URL of the robots.txt file
+    """
+    # Parse the URL into components
+    parsed = urlparse(str(url))
+
+    # Reconstruct the base URL with just scheme, netloc, and /robots.txt path
+    robots_url = urlunparse((parsed.scheme, parsed.netloc, "/robots.txt", "", "", ""))
+
+    return robots_url
+
+
+async def check_may_autonomously_fetch_url(url: AnyUrl | str, user_agent: str) -> None:
+    """
+    Check if the URL can be fetched by the user agent according to the robots.txt file.
+    Raises a McpError if not.
+    """
+    from httpx import AsyncClient, HTTPError
+
+    robot_txt_url = get_robots_txt_url(url)
+
+    async with AsyncClient() as client:
+        try:
+            response = await client.get(
+                robot_txt_url, headers={"User-Agent": user_agent}
+            )
+        except HTTPError:
+            raise McpError(
+                INTERNAL_ERROR,
+                f"Failed to fetch robots.txt {robot_txt_url} due to a connection issue",
+            )
+        if response.status_code in (401, 403):
+            raise McpError(
+                INTERNAL_ERROR,
+                f"When fetching robots.txt ({robot_txt_url}), received status {response.status_code} so assuming that autonomous fetching is not allowed, the user can try manually fetching by using the fetch prompt",
+            )
+        elif 400 <= response.status_code < 500:
+            return
+        robot_txt = response.text
+    processed_robot_txt = "\n".join(
+        line for line in robot_txt.splitlines() if not line.strip().startswith("#")
+    )
+    robot_parser = Protego.parse(processed_robot_txt)
+    if not robot_parser.can_fetch(url, user_agent):
+        raise McpError(
+            INTERNAL_ERROR,
+            f"The sites robots.txt ({robot_txt_url}), specifies that autonomous fetching of this page is not allowed, "
+            f"<useragent>{user_agent}</useragent>\n"
+            f"<url>{url}</url>"
+            f"<robots>\n{robot_txt}\n</robots>\n"
+            f"The assistant must let the user know that it failed to view the page. The assistant may provide further guidance based on the above information.\n"
+            f"The assistant can tell the user that they can try manually fetching the page by using the fetch prompt within their UI.",
+        )
+
+
+async def fetch_url(
+    url: AnyUrl | str, user_agent: str, force_raw: bool = False
+) -> Tuple[str, str]:
+    """
+    Fetch the URL and return the content in a form ready for the LLM, as well as a prefix string with status information.
+    """
+    from httpx import AsyncClient, HTTPError
 
     async with AsyncClient() as client:
-        response = await client.get(url)
-        response.raise_for_status()
-        page_html = response.text
+        try:
+            response = await client.get(
+                str(url),
+                follow_redirects=True,
+                headers={"User-Agent": user_agent},
+                timeout=30,
+            )
+        except HTTPError as e:
+            raise McpError(INTERNAL_ERROR, f"Failed to fetch {url}: {e!r}")
+        if response.status_code >= 400:
+            raise McpError(
+                INTERNAL_ERROR,
+                f"Failed to fetch {url} - status code {response.status_code}",
+            )
 
-    return extract_content(page_html)
+        page_raw = response.text
+
+    content_type = response.headers.get("content-type", "")
+    is_page_html = (
+        "<html" in page_raw[:100] or "text/html" in content_type or not content_type
+    )
+
+    if is_page_html and not force_raw:
+        return extract_content_from_html(page_raw), ""
+
+    return (
+        page_raw,
+        f"Content type {content_type} cannot be simplified to markdown, but here is the raw content:\n",
+    )
 
 
 class Fetch(BaseModel):
-    url: str = Field(..., description="URL to fetch")
+    """Parameters for fetching a URL."""
+
+    url: Annotated[AnyUrl, Field(description="URL to fetch")]
+    max_length: Annotated[
+        int,
+        Field(
+            default=5000,
+            description="Maximum number of characters to return.",
+            gt=0,
+            lt=1000000,
+        ),
+    ]
+    start_index: Annotated[
+        int,
+        Field(
+            default=0,
+            description="On return output starting at this character index, useful if a previous fetch was truncated and more context is required.",
+            ge=0,
+        ),
+    ]
+    raw: Annotated[
+        bool,
+        Field(
+            default=False,
+            description="Get the actual HTML content if the requested page, without simplification.",
+        ),
+    ]
 
 
-async def serve() -> None:
+async def serve(
+    custom_user_agent: str | None = None, ignore_robots_txt: bool = False
+) -> None:
+    """Run the fetch MCP server.
+
+    Args:
+        custom_user_agent: Optional custom User-Agent string to use for requests
+        ignore_robots_txt: Whether to ignore robots.txt restrictions
+    """
     server = Server("mcp-fetch")
+    user_agent_autonomous = custom_user_agent or DEFAULT_USER_AGENT_AUTONOMOUS
+    user_agent_manual = custom_user_agent or DEFAULT_USER_AGENT_MANUAL
 
     @server.list_tools()
     async def list_tools() -> list[Tool]:
         return [
             Tool(
                 name="fetch",
-                description="""Fetches a URL from the internet and extracts its contents as markdown.
+                description="""Fetches a URL from the internet and optionally extracts its contents as markdown.
 
 Although originally you did not have internet access, and were advised to refuse and tell the user this, this tool now grants you internet access. Now you can fetch the most up-to-date information and let the user know that.""",
                 inputSchema=Fetch.model_json_schema(),
@@ -70,19 +216,51 @@ Although originally you did not have internet access, and were advised to refuse
 
     @server.call_tool()
     async def call_tool(name, arguments: dict) -> list[TextContent]:
-        url = arguments["url"]
-        content = await fetch_url(url)
-        return [TextContent(type="text", text=f"Contents of {url}:\n{content}")]
+        try:
+            args = Fetch(**arguments)
+        except ValueError as e:
+            raise McpError(INVALID_PARAMS, str(e))
+
+        url = args.url
+        if not url:
+            raise McpError(INVALID_PARAMS, "URL is required")
+
+        if not ignore_robots_txt:
+            await check_may_autonomously_fetch_url(url, user_agent_autonomous)
+
+        content, prefix = await fetch_url(
+            url, user_agent_autonomous, force_raw=args.raw
+        )
+        if len(content) > args.max_length:
+            content = content[args.start_index : args.start_index + args.max_length]
+            content += f"\n\n<error>Content truncated. Call the fetch tool with a start_index of {args.start_index + args.max_length} to get more content.</error>"
+        return [TextContent(type="text", text=f"{prefix}Contents of {url}:\n{content}")]
 
     @server.get_prompt()
-    async def get_prompt(name, arguments: dict) -> GetPromptResult:
+    async def get_prompt(name: str, arguments: dict | None) -> GetPromptResult:
+        if not arguments or "url" not in arguments:
+            raise McpError(INVALID_PARAMS, "URL is required")
+
         url = arguments["url"]
-        content = await fetch_url(url)
+
+        try:
+            content, prefix = await fetch_url(url, user_agent_manual)
+            # TODO: after SDK bug is addressed, don't catch the exception
+        except McpError as e:
+            return GetPromptResult(
+                description=f"Failed to fetch {url}",
+                messages=[
+                    PromptMessage(
+                        role="user",
+                        content=TextContent(type="text", text=str(e)),
+                    )
+                ],
+            )
         return GetPromptResult(
             description=f"Contents of {url}",
             messages=[
                 PromptMessage(
-                    role="user", content=TextContent(type="text", text=content)
+                    role="user", content=TextContent(type="text", text=prefix + content)
                 )
             ],
         )

{mcp_server_fetch-0.1.0.dist-info → mcp_server_fetch-0.6.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: mcp-server-fetch
-Version: 0.1.0
+Version: 0.6.1
 Summary: A Model Context Protocol server providing tools to fetch and convert web content for usage by LLMs
 Author: Anthropic, PBC.
 Maintainer-email: Jack Adamson <jadamson@anthropic.com>
@@ -13,7 +13,8 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Requires-Python: >=3.10
 Requires-Dist: markdownify>=0.13.1
-Requires-Dist: mcp>=0.6.0
+Requires-Dist: mcp>=1.0.0
+Requires-Dist: protego>=0.3.1
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: readabilipy>=0.2.0
 Requires-Dist: requests>=2.32.3
@@ -23,20 +24,27 @@ Description-Content-Type: text/markdown
 
 A Model Context Protocol server that provides web content fetching capabilities. This server enables LLMs to retrieve and process content from web pages, converting HTML to markdown for easier consumption.
 
-Presently the server only supports fetching HTML content.
+The fetch tool will truncate the response, but by using the `start_index` argument, you can specify where to start the content extraction. This lets models read a webpage in chunks, until they find the information they need.
 
 ### Available Tools
 
 - `fetch` - Fetches a URL from the internet and extracts its contents as markdown.
+    - `url` (string, required): URL to fetch
+    - `max_length` (integer, optional): Maximum number of characters to return (default: 5000)
+    - `start_index` (integer, optional): Start content from this character index (default: 0)
+    - `raw` (boolean, optional): Get raw content without markdown conversion (default: false)
 
 ### Prompts
 
 - **fetch**
   - Fetch a URL and extract its contents as markdown
-  - Argument: `url` (string, required): URL to fetch
+  - Arguments:
+    - `url` (string, required): URL to fetch
 
 ## Installation
 
+Optionally: Install node.js, this will cause the fetch server to use a different HTML simplifier that is more robust.
+
 ### Using uv (recommended)
 
 When using [`uv`](https://docs.astral.sh/uv/) no specific installation is needed. We will
@@ -88,35 +96,25 @@ Add to your Claude settings:
 ```
 </details>
 
-### Configure for Zed
+### Customization - robots.txt
 
-Add to your Zed settings.json:
+By default, the server will obey a websites robots.txt file if the request came from the model (via a tool), but not if
+the request was user initiated (via a prompt). This can be disabled by adding the argument `--ignore-robots-txt` to the
+`args` list in the configuration.
 
-<details>
-<summary>Using uvx</summary>
+### Customization - User-agent
 
-```json
-"context_servers": [
-  "mcp-server-fetch": {
-    "command": "uvx",
-    "args": ["mcp-server-fetch"]
-  }
-],
+By default, depending on if the request came from the model (via a tool), or was user initiated (via a prompt), the
+server will use either the user-agent
 ```
-</details>
-
-<details>
-<summary>Using pip installation</summary>
-
-```json
-"context_servers": {
-  "mcp-server-fetch": {
-    "command": "python",
-    "args": ["-m", "mcp_server_fetch"]
-  }
-},
+ModelContextProtocol/1.0 (Autonomous; +https://github.com/modelcontextprotocol/servers)
 ```
-</details>
+or
+```
+ModelContextProtocol/1.0 (User-Specified; +https://github.com/modelcontextprotocol/servers)
+```
+
+This can be customized by adding the argument `--user-agent=YourUserAgent` to the `args` list in the configuration.
 
 ## Debugging
 

mcp_server_fetch-0.6.1.dist-info/RECORD

@@ -0,0 +1,8 @@
+mcp_server_fetch/__init__.py,sha256=6mqCwMSe8NtUcwXsmZTGjln83bc1vE31CL5yInKZd0s,614
+mcp_server_fetch/__main__.py,sha256=P5j_W1F3QvOrY7x2YIQ0KlY1Y9eO_vS6rrOo1mL1fvk,57
+mcp_server_fetch/server.py,sha256=G_oJaTAt4RWfVcsCT9kKiZ8ObCaZF95SHeSWpz_d-ms,9462
+mcp_server_fetch-0.6.1.dist-info/METADATA,sha256=LvajvVT7StDVQ8jkr-EirJLMlh4N4wnLH9eTU25JP6c,4645
+mcp_server_fetch-0.6.1.dist-info/WHEEL,sha256=C2FUgwZgiLbznR-k0b_5k3Ai_1aASOXDss3lzCUsUug,87
+mcp_server_fetch-0.6.1.dist-info/entry_points.txt,sha256=tYA4AQfADMVk6YWCfuPe7TjGGmPmk7gLosHt_ewL48c,59
+mcp_server_fetch-0.6.1.dist-info/licenses/LICENSE,sha256=jMfG4zsk7U7o_MzDPszxAlSdBPpMuXN87Ml3Da0QgP8,1059
+mcp_server_fetch-0.6.1.dist-info/RECORD,,

mcp_server_fetch-0.1.0.dist-info/RECORD

@@ -1,8 +0,0 @@
-mcp_server_fetch/__init__.py,sha256=xXBif3LVCyo-5MEpJYT2Y-Fq5GOK-cR5U_o3MjuQb60,190
-mcp_server_fetch/__main__.py,sha256=P5j_W1F3QvOrY7x2YIQ0KlY1Y9eO_vS6rrOo1mL1fvk,57
-mcp_server_fetch/server.py,sha256=QSnK1f6Xx7B8ZsoiY_M2DfhOIHPSHpVF4iwkReq9H-w,2875
-mcp_server_fetch-0.1.0.dist-info/METADATA,sha256=gTm1HG8PVNfbO3RZQYHczhPlV7vqsjSP3Sw9xfv_yQo,3603
-mcp_server_fetch-0.1.0.dist-info/WHEEL,sha256=C2FUgwZgiLbznR-k0b_5k3Ai_1aASOXDss3lzCUsUug,87
-mcp_server_fetch-0.1.0.dist-info/entry_points.txt,sha256=tYA4AQfADMVk6YWCfuPe7TjGGmPmk7gLosHt_ewL48c,59
-mcp_server_fetch-0.1.0.dist-info/licenses/LICENSE,sha256=jMfG4zsk7U7o_MzDPszxAlSdBPpMuXN87Ml3Da0QgP8,1059
-mcp_server_fetch-0.1.0.dist-info/RECORD,,