Bright Data MCP
@brightdata
A powerful Model Context Protocol (MCP) server that provides an all-in-one solution for public web access.
概览
What is Bright Data MCP?
Bright Data MCP is an MCP server that gives AI assistants real-time web capabilities, enabling them to search, scrape, and access live data without getting blocked or rate-limited. It is built by Bright Data, offers a free tier of 5,000 credits per month, and works with any LLM.
How to use Bright Data MCP?
You can use the hosted server at https://mcp.brightdata.com/mcp?token=YOUR_API_TOKEN (zero installation) or run it locally via npx @brightdata/mcp with the API_TOKEN environment variable set. Configure tool groups and modes using environment variables like GROUPS or PRO_MODE in your MCP client’s JSON settings.
Key features of Bright Data MCP
- Smart web search optimized for AI consumption
- Clean markdown content extraction from any page
- Global access bypassing geo-restrictions automatically
- Anti-bot protection preventing blocks and rate limits
- Browser automation for controlling real browsers (Pro mode)
- Over 60+ structured data tools across multiple domains
Use cases of Bright Data MCP
- Real-time research on prices, news, and live data
- E-commerce intelligence for product and price monitoring
- AI agents that browse the web autonomously
- Coding agents looking up npm and PyPI package info
- GEO and brand visibility tracking across LLMs
FAQ from Bright Data MCP
What is included in the free tier?
The free Rapid mode gives you 5,000 credits per month (auto-renewed, no credit card required) and includes web search, scraping with web unlocker, and AI-ranked Discover search. Credits are consumed per request.
How do I get an API token?
You need to sign up for a Bright Data account to obtain an API token. The token is passed via the API_TOKEN environment variable or in the hosted server URL.
Does Bright Data MCP work with any LLM?
Yes, it works with Claude, GPT, Gemini, Llama, and any MCP-compatible agent such as Claude Code, Cursor, and Windsurf.
What are the base tools always available?
The base tools always enabled are search_engine, search_engine_batch, scrape_as_markdown, scrape_batch, and discover. These are part of the default Rapid mode.
Can I customize which tools are available?
Yes. You can use the GROUPS environment variable to enable curated tool bundles (e.g., ecommerce, browser, code) and the TOOLS variable to add individual tools on top. Setting PRO_MODE=true enables all tools.
