Hyperbrowser MCP Server

Overview

Hyperbrowser MCP Server connects MCP clients to Hyperbrowser's hosted browser automation platform. The upstream README describes tools for scraping pages, crawling websites, extracting structured data, searching with Bing, managing persistent browser profiles, and running browser agents through Browser Use, OpenAI computer use, or Claude computer use.

Use it when an agent needs a real browser-backed workflow rather than a simple HTTP fetch: dynamic pages, multi-page crawl jobs, structured extraction from web content, or controlled browser-agent runs.

Source Review

This listing is grounded in:

• https://github.com/hyperbrowserai/mcp
• https://github.com/hyperbrowserai/mcp/blob/main/README.md
• https://github.com/hyperbrowserai/mcp/blob/main/package.json
• https://github.com/hyperbrowserai/mcp/blob/main/src/server.ts
• https://github.com/hyperbrowserai/mcp/blob/main/LICENSE
• https://registry.npmjs.org/hyperbrowser-mcp/latest

The repository README identifies the project as Hyperbrowser's Model Context Protocol server. The npm package is published as hyperbrowser-mcp, exposes a hyperbrowser-mcp binary, and requires Node.js 18 or newer.

Install

For Claude Code, add the server with an API key:

claude mcp add hyperbrowser \
  -e HYPERBROWSER_API_KEY=your-api-key \
  -- npx -y hyperbrowser-mcp

For JSON-based MCP clients, use:

{
  "mcpServers": {
    "hyperbrowser": {
      "command": "npx",
      "args": ["-y", "hyperbrowser-mcp"],
      "env": {
        "HYPERBROWSER_API_KEY": "YOUR-API-KEY"
      }
    }
  }
}

The upstream README also documents Cursor, Windsurf, Claude Desktop, local development, and Smithery installation examples.

Tool Scope

The README lists these tools:

Operating Guidance

• Use read-only scraping, crawling, extraction, or search before browser-agent actions that click, type, submit forms, or change account state.
• Scope browser-agent tasks to explicit domains, accounts, and goals.
• Use temporary profiles for one-off tasks and delete them after use.
• Keep persistent profiles separate by environment, customer, and permission boundary.
• Review extracted data before using it in reports, code generation, outreach, compliance tasks, or customer-facing output.
• Confirm terms of service, authentication requirements, and rate limits before scraping or crawling third-party sites.

Best Use Cases

• Scrape dynamic docs pages for implementation context when ordinary fetches miss rendered content.
• Crawl approved product or documentation pages into Markdown for agent review.
• Extract structured data from messy HTML tables, profiles, directories, or listings.
• Run controlled browser-agent tasks that need visual page state or multi-step navigation.
• Keep persistent browser profiles for repeated testing against approved non-production accounts.

Duplicate Review

Checked current MCP, agents, skills, and open pull requests for hyperbrowserai/mcp, hyperbrowser-mcp, Hyperbrowser MCP, Hyperbrowser browser automation MCP, and related browser-automation entries. Existing Browserbase, BrowserMCP, Chrome MCP, Playwright MCP, Firecrawl, Exa, Jina, and scraping entries cover adjacent workflows, but no dedicated Hyperbrowser MCP Server entry or matching source URL was found.

Disclosure

Editorial listing. No paid placement or affiliate link is used. Hyperbrowser MCP is published by Hyperbrowser and the upstream repository uses MIT license text; the npm package metadata currently reports ISC, so verify license metadata if you redistribute bundled artifacts.