What is MCP?
Model Context Protocol (MCP) is the standard way AI agents interact with external tools. Instead of building custom HTTP integrations, your agent discovers and uses tools through a unified protocol.
SearchX provides an MCP server with 4 search tools out of the box.
Setup (5 minutes)
Claude Code / Claude Desktop:
``json
{
"mcpServers": {
"searchx": {
"command": "python3",
"args": ["path/to/mcp-search-server/server.py"],
"env": {
"SEARCH_API_URL": "https://searchx.dev"
}
}
}
}
``
Available Tools
search General web search with mode selection. ``` search("kubernetes pod networking", mode="hybrid", agent="coder") ```
search_for_code Optimized for code-related queries. Prioritizes docs, GitHub, Stack Overflow. ``` search_for_code("golang goroutines tutorial") ```
search_for_research Optimized for academic and research queries. ``` search_for_research("transformer architecture attention mechanism") ```
search_suggest Autocomplete suggestions. ``` search_suggest("kube") → ["Kubernetes", "Kubernetes Pods", "Kubernetes Service"] ```
Why MCP > Direct API
MCP lets the agent decide when to search. Instead of hardcoding search into your pipeline, the agent autonomously chooses to search when it needs information. This leads to more natural, efficient tool use.