crawl_results is the parameter that changes how this gets used: rather than a search followed by one fetch per promising result, the top hits come back with their content already pulled, and each successful crawl adds 1 credit to the base 1-credit request. The hosted endpoint means nothing to install for web clients, and because discovery answers without a credential you can see the tool surface before deciding to sign in.
The official MCP server for Search1API: web search across a wide set of engines and verticals, news search, single-page crawling, sitemap discovery, and trending topics. Tool discovery is answered without a credential, so a client can inspect the server before anyone signs in.
- search — web results with a citable id, title, and url, over google, bing, duckduckgo, yahoo, x, reddit, github, youtube, arxiv, wechat, bilibili, imdb, and wikipedia
- news — the same shape against google, bing, duckduckgo, yahoo, and hackernews
- crawl — the full content of a URL, including one that came back from a search result
- sitemap — every related link discoverable from a URL
- trending — what is currently rising on github or hackernews
- crawl_results on search and news, which pulls full page content for the top hits inline instead of leaving the agent to fetch them
- Include and exclude site lists, and a time_range of day, month, or year
A Search1API credential for tool calls; discovery works without one. The hosted endpoint is https://mcp.search1api.com/mcp over Streamable HTTP — connect with OAuth 2.1 where your client supports it, or send an Authorization Bearer header with a key from the Search1API dashboard. The URL query-parameter form still works but is the legacy path, and it puts the key into URLs, logs, and shell history. To run it locally instead, use Node.js 20 or newer with npx -y search1api-mcp and SEARCH1API_KEY set. Self-hosted HTTP deployments behind a proxy add their internal hostnames to MCP_ALLOWED_HOSTS, and browser-based clients add their origins to MCP_ALLOWED_ORIGINS.
One command plus a key — claude mcp add --transport http search1api https://mcp.search1api.com/mcp, then supply credentials