Bright Data MCP

Add to your client

Copy the config for your MCP client and paste it into its config file.

{
  "mcpServers": {
    "bright-data-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@brightdata/mcp"
      ],
      "env": {
        "API_TOKEN": "<your-brightdata-api-token>"
      }
    }
  }
}

Before you start

• Node.js installed (for local npx execution); not needed if you use the hosted remote server
• A Bright Data API token — sign up at https://brightdata.com and create the token in your account (free tier available, no credit card required)
• Set the token via the API_TOKEN environment variable (local) or the ?token= query param (hosted)

About Bright Data MCP

Tools & capabilities (12)

When to use it

• Use it when your agent needs current, real-time web content that ordinary fetches get blocked on.
• Use it when you want clean markdown of arbitrary pages (Web Unlocker) instead of raw, bot-blocked HTML.
• Use it when you need structured records from specific sites — Amazon, LinkedIn, Zillow, Crunchbase, app stores — via the web_data_* tools.
• Use it when a task requires real browser interaction (click, screenshot, snapshot) behind anti-bot defenses, via the cloud Scraping Browser.
• Use it when you want geo-restriction and CAPTCHA handling done for you across LLMs (Claude, GPT, Gemini).
• Use it when you want zero local setup — point a client at the hosted remote server with just a token.

Quick setup

1. 1Create a Bright Data account at https://brightdata.com and generate an API token (free tier, no card).
2. 2For hosted: add the URL https://mcp.brightdata.com/mcp?token=YOUR_TOKEN to your MCP client. For local: configure command "npx" with args ["@brightdata/mcp"] and API_TOKEN in env.
3. 3Optionally set PRO_MODE=true to enable all 60+ tools, or use GROUPS / TOOLS to enable subsets.
4. 4Restart your MCP client (Claude Desktop, Cursor, Windsurf, etc.).
5. 5Verify with a simple search_engine or scrape_as_markdown call.

Bright Data MCP FAQ