Hyperbrowser MCP Server

npx -y hyperbrowser-mcp

{
  "mcpServers": {
    "hyperbrowser-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "hyperbrowser-mcp"
      ]
    }
  }
}

Hyperbrowser runs headless sessions on its own infrastructure, so the machine holding the agent does not need Chrome, proxies or stealth patches. The server exposes scraping a page, crawling a site to a depth, structured extraction against a schema, and delegating a whole task to a browser agent. For long research loops the appeal is operational rather than technical — sessions that get blocked or hang are the provider's problem, not a stuck subprocess on a laptop.