Summary
Search that works the minute you paste the config.
Most search servers stop at a signup page; this one runs `uvx heventure-search-mcp` and starts returning results. web_search takes an engine choice and a result count per call, and get_page_content turns any hit into text, which is the whole look-it-up-then-read-it loop.
What it is
A Python server that searches DuckDuckGo, Bing and Google without an account, and pulls the readable text out of any page it finds.
What you get
- Web search across DuckDuckGo, Bing and Google — one engine or several at once, chosen per call
- A result count you choose per search, within the range the API allows
- Readable text extracted from any page URL
- Async requests with caching, so repeat queries do not refetch
- SerpAPI and Tavily keys accepted if you want their results, with nothing breaking when they are absent
Requirements
`pip install heventure-search-mcp`, or run it straight from a client config with uvx. No API key and no account — the default engines are the free ones. A Docker image serves it on port 8080.
Setup effort
One command plus a key — pip install heventure-search-mcp, then supply credentials