What it does
The Oxylabs MCP server bridges Claude and the live web, exposing two sets of tools for scraping and data extraction. Through the Web Scraper API, it handles general-purpose URL scraping plus specialized extractors for Google Search and Amazon product pages. The AI Studio tools use Claude-powered extraction to scrape arbitrary URLs into structured JSON or Markdown, crawl multi-page sites, navigate sites via natural-language browser control, and search the web. The server handles JavaScript rendering, anti-scraping defenses, and geo-restricted content from 195+ countries.
Who it's for
AI developers building Claude applications that need current web data, market researchers using LLMs to monitor competitors or pricing, teams automating workflows like price comparison or market intelligence gathering, and data analysts training or prompting models on web-sourced information.
Common use cases
- Scrape live Google Search or Amazon results to feed Claude with current market pricing and product data
- Crawl multiple pages of a website and extract consolidated data in JSON or Markdown format
- Control a browser programmatically based on natural-language instructions to perform complex navigation tasks
- Bypass anti-scraping defenses and geo-blocking to access restricted content from 195+ countries
- Extract structured data from JavaScript-heavy sites that traditional static scrapers cannot handle
Setup pitfalls
- Requires Oxylabs credentials (free trial available): either Web Scraper API username/password or AI Studio API key. Leaving placeholder values will expose tools that do not function.
- Credentials must be stored in environment variables; the server will expose only tools corresponding to provided credentials—if both sets are supplied, both tool sets become available.
- The server reads filesystem paths, so ensure file access is restricted appropriately when deployed.