brightdata-mcp: an MCP server providing AI assistants with unblocked real-time web search and structured data extraction

What it solves

This project provides a Model Context Protocol (MCP) server that allows AI assistants to access real-time web data without being blocked by CAPTCHAs, rate limits, or geo-restrictions. It eliminates the limitation of LLMs having outdated training data by giving them a reliable way to browse and extract information from the live web.

How it works

It acts as a bridge between an MCP-compatible AI client (like Claude Desktop, Cursor, or Windsurf) and Bright Data's web data platform. The server can be run locally via npx or used as a hosted service. It exposes a variety of tools to the LLM, ranging from basic web search and markdown extraction to specialized datasets (e.g., e-commerce, social media, finance) and full browser automation.

Who it’s for

• AI Agent Developers: Those building agents that need real-time web browsing capabilities.
• Coding Agents: Developers using tools like Claude Code or Cursor who need real-time package metadata from npm and PyPI.
• Researchers and Analysts: People performing real-time market analysis, e-commerce intelligence, or academic research.
• Brand Managers: Users wanting to monitor how different LLMs (ChatGPT, Grok, Perplexity) perceive their brand (GEO).

Highlights

• Enterprise-grade Unblocking: Built-in capabilities to bypass anti-bot protections and geo-restrictions.
• AI-Ready Content: Extracts webpages as clean Markdown for better LLM consumption.
• Specialized Tool Groups: Curated bundles for specific domains like e-commerce, social media, finance, and coding package intelligence.
• Flexible Deployment: Supports both local installation and a zero-config hosted server.
• Free Tier: Offers 5,000 free credits per month for prototyping.