Skip to content

Tool suggestion: anybrowse for web scraping with Cloudflare bypass #2109

@kc23go

Description

@kc23go

Problem with current web tools

smolagents VisitWebpageTool fails on Cloudflare-protected sites. For research agents this silently degrades quality -- empty content on a large fraction of useful URLs.

Alternative: anybrowse

anybrowse runs real residential Chrome and returns clean markdown. Works on Cloudflare-protected sites.

import requests
from smolagents import tool

@tool
def scrape_url(url: str) -> str:
    """Scrape any URL and return clean markdown content, including Cloudflare-protected sites."""
    r = requests.post("https://anybrowse.dev/scrape", json={"url": url})
    if r.status_code == 200:
        return r.json().get("markdown", "")
    return f"Failed to scrape: HTTP {r.status_code}"

Or via MCP:

{
  "mcpServers": {
    "anybrowse": {
      "type": "streamable-http",
      "url": "https://anybrowse.dev/mcp"
    }
  }
}

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions