Summary
It keeps searching when one engine starts blocking you, and tells you which one did.
search_web queries several engines with automatic fallback, and the state that fallback depends on is exposed rather than hidden: get_search_engine_status names the active browser backend and marks each engine available or blocked, and reset_search_engines clears that list once they recover. get_webpage_content then pulls the text behind a result, with max_length to cap what comes back. The 4 tools stop at text — there is no crawling, no structured extraction and no format conversion.
What it is
An MCP server for web search over several engines with fallback, plus fetching of page text.
What you get
- Search across multiple engines that falls back automatically when one of them stops answering, with snippets and a result cap you set.
- Visibility into that fallback — the active browser backend and each engine's availability or blocked state — and a reset that clears the blocked list.
- Page text fetched from a URL, capped at a length you choose.
Requirements
Nothing beyond the machine it runs on — no account, no key, no environment variable.
Setup effort
One command — uvx git+https://github.com/pranavms13/web-search-mcp