MCP Server LIVE

Search the web from any AI agent via Model Context Protocol. Keyword, semantic, and hybrid search across 1M+ pages.

Endpoints

POST https://mcp.searchx.dev/mcp          # Streamable HTTP transport
GET  https://mcp.searchx.dev/mcp          # SSE transport
GET  https://mcp.searchx.dev/mcp/health   # Health check
GET  https://mcp.searchx.dev/mcp/docs     # This page

Tools (6)

search General web search — keyword, semantic, or hybrid mode

search_suggest Autocomplete suggestions

search_for_code Code, docs, APIs, libraries — prioritizes GitHub, StackOverflow

search_for_research Papers, data, facts — prioritizes arxiv, academic sources

search_for_writing Content sources, references, articles for writing

search_for_market Market trends, competitors, SEO data

Quick Start

# Initialize
curl -X POST https://mcp.searchx.dev/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"my-agent","version":"1"}},"id":1}'

# Search
curl -X POST https://mcp.searchx.dev/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"search","arguments":{"query":"kubernetes deployment","mode":"hybrid"}},"id":2}'

Claude Code / Cursor

# Option 1: Direct HTTP (no install needed)
# Just use the endpoint: https://mcp.searchx.dev/mcp

# Option 2: Python MCP client (standalone)
{
  "mcpServers": {
    "searchx": {
      "command": "python3",
      "args": ["mcp-search-server/server.py"],
      "env": {"SEARCH_API_URL": "https://searchx.dev"}
    }
  }
}

Response Format

{
  "results": [{
    "title": "Page Title",
    "url": "https://...",
    "snippet": "Clean text with **highlights**...",
    "trust_score": 0.92,
    "citation": "[Title](url) — domain",
    "metadata": {"source_type": "documentation"}
  }]
}

● Server online — health check