home / mcp / seo crawler mcp server

SEO Crawler MCP Server

Provides a local, SQLite-backed MCP server to crawl websites and analyze technical SEO data with AI-assisted insights.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "houtini-ai-seo-crawler-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@houtini/seo-crawler-mcp"
      ],
      "env": {
        "DEBUG": "false",
        "OUTPUT_DIR": "C:\\\\seo-audits"
      }
    }
  }
}

You can run the SEO Crawler MCP server locally to crawl websites, store crawl data in an embedded SQLite database, and analyze the results with AI-assisted insights. This MCP server orchestrates crawling, data storage, and SQL-driven analysis, enabling you to run big site audits entirely on your machine or in your trusted environment without external API calls.

How to use

To use the SEO Crawler MCP server, you connect your MCP client or Claude-like assistant to the local MCP instance. You first start a crawl to gather pages, links, and metadata, then run a structured analysis on the collected data. You can later query specific SEO issues, review summaries, or request detailed reports. This workflow supports large sites by performing crawls in the background and letting your AI assistant extract actionable insights.

How to install

Prerequisites: you need Node.js and npm installed on your system. You should also have a working Claude Desktop or MCP client configured to communicate with an MCP server.

Step 1. Quick Install via NPX (run from any terminal) start the server using the NPX-based configuration.

{
  "mcpServers": {
    "seo_crawler": {
      "command": "npx",
      "args": ["-y", "@houtini/seo-crawler-mcp"],
      "env": {
        "OUTPUT_DIR": "C:\\seo-audits"
      }
    }
  }
}

Restart your MCP client after adding the configuration. Four tools will be available for interactions with the crawl and analysis stack. You can run crawls and then analyze results directly through the client.

# Development Install
cd C:\MCP\seo-crawler-mcp
npm install
npm run build

# Then use a local runtime command

Additional installation notes

Alternatively, you can run the server from a local build using Node.js and the built index. This is useful for development and testing.

{
  "mcpServers": {
    "seo_crawler_dev": {
      "command": "node",
      "args": ["C:\\MCP\\seo-crawler-mcp\\build\\index.js"],
      "env": {
        "OUTPUT_DIR": "C:\\seo-audits",
        "DEBUG": "false"
      }
    }
  }
}

Usage patterns and commands

- Start a crawl from your MCP client for a site like https://example.com with default settings. The crawl runs asynchronously and stores data locally in SQLite.

- Once the crawl finishes, you can trigger a detailed SEO analysis that focuses on critical issues, security concerns, and optimization opportunities.

Available tools

run_seo_audit

Crawl a website and extract comprehensive SEO data into SQLite.

analyze_seo

Run comprehensive SEO analysis on a completed crawl.

query_seo_data

Execute specific SEO queries by name.

list_seo_queries

Discover available SEO analysis queries.