Summary of an article published on The Web Scraping Club on 2026-07-26 (free to read). Read the full article

SearchApi has released an official remote MCP server called Web Search MCP. It gives AI clients one interface to the same endpoints as its scraping APIs, covering Google, Bing, YouTube, Amazon, Airbnb, TikTok and more. For anyone building agents that need real-time search data, the question is whether it connects cleanly and handles anti-bot protections that general-purpose fetch tools struggle with.

I cover what SearchApi is, how the server works, and what it costs. It is remote-only, runs over Streamable HTTP and supports OAuth 2.0. Then I set it up in Claude Code and ran a sentiment analysis workflow on restaurant reviews from Google Maps and Tripadvisor. I walk through creating the integration, picking the tools, connecting the client, checking the connection and running the prompt. I also inspect the raw output of each tool call and list other use cases. I close with my verdict after testing it.

What’s inside

  • The server exposes over 80 tools, and you choose which ones go into each integration, either from recommended bundles or by picking your own.
  • Review tools need an identifier, so the workflow also needs search tools to find place IDs and URLs first.
  • Each MCP tool call counts as one API request under the pay-per-success model, with a cap of 10 integrations per account and standard rate limits.
  • Other use cases include travel price comparison, ad monitoring, e-commerce research, news tracking and Google Scholar queries.

In the full article: The step-by-step setup in the SearchApi dashboard and Claude Code, the exact prompt, the screenshots of each tool call’s output, the pricing figures and my final verdict on the server.

Read the full article on The Web Scraping Club