Summary
Search with nothing to sign up for — and the limits that come with it.
Scraping rather than an API is what removes the key, and it is also what constrains it: Google may temporarily block you if searches come too fast, results occasionally arrive without descriptions when the page structure shifts, and complex search operators may not behave as expected. Keep the frequency modest and the limit small, and treat it as a personal-use tool rather than something to put behind a product.
What it is
A Node MCP server that performs a web search by scraping Google's results and returning them as structured records, with no search API behind it.
What you get
- `search` — takes a `query` and an optional `limit` (default 5, maximum 10)
- Structured results: title, url and description per hit, rather than a wall of text
- No keys and no authentication to arrange before the first call
Requirements
Clone the repo, `npm install`, `npm run build`, then point your client at the built build/index.js. Works the same in VS Code with the Claude Dev extension or in Claude Desktop.