mcp_websearch/mcp_server.py
2026-10-13 14:20:24 +02:00

74 lines
2.6 KiB
Python

import os
import asyncio
import httpx
import trafilatura
from fastmcp import FastMCP
MOBILE_USER_AGENT = (
"Mozilla/5.0 (iPhone; CPU iPhone OS 17_5 like Mac OS X) "
"AppleWebKit/605.1.15 (KHTML, like Gecko) "
"Version/17.5 Mobile/15E148 Safari/604.1"
)
# Disable DNS rebinding protection to allow remote tunnel headers
mcp = FastMCP("Mobile Web MCP Server")
@mcp.tool()
async def search_duckduckgo_lite(query: str) -> str:
"""Searches DuckDuckGo Lite using a mobile User-Agent."""
headers = {
"User-Agent": MOBILE_USER_AGENT,
"Content-Type": "application/x-www-form-urlencoded",
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
"Accept-Language": "en-US,en;q=0.9",
}
async with httpx.AsyncClient(follow_redirects=True) as client:
try:
response = await client.post(
"https://lite.duckduckgo.com/lite/",
data={"q": query},
headers=headers,
timeout=15.0
)
response.raise_for_status()
text = response.text
cleaned_content = trafilatura.extract(text, output_format="markdown", include_links=True, include_images=False)
if len(cleaned_content) > 6000:
return cleaned_content[:6000]
return cleaned_content or "Fail to extract"
except Exception as e:
return f"Failed to search: {str(e)}"
@mcp.tool()
async def fetch_html(url: str) -> str:
"""Fetches raw HTML content from a URL using a spoofed mobile User-Agent."""
headers = {
"User-Agent": MOBILE_USER_AGENT,
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
"Accept-Language": "en-US,en;q=0.9",
}
async with httpx.AsyncClient(follow_redirects=True) as client:
try:
response = await client.get(url, headers=headers, timeout=15.0)
response.raise_for_status()
text = response.text
cleaned_content = trafilatura.extract(text, output_format="markdown", include_links=True, include_images=False, url=url)
if len(cleaned_content) > 6000:
return cleaned_content[:6000]
return cleaned_content or "Fail to extract"
except Exception as e:
return f"Failed to fetch URL: {str(e)}"
if __name__ == "__main__":
#print(asyncio.run(fetch_html("https://www.google.com")))
#print(asyncio.run(search_duckduckgo_lite("bornhack")))
mcp.run(
transport="http",
host="127.0.0.1",
port=8005,
path="/mcp",
allowed_origins=["*"],
allowed_hosts=["*"]
)