Six tools
| Tool | What it does |
|---|---|
audit_url | Full audit of one URL: score, engine readiness, gates, category scores, prioritized fixes. Summary, full or raw JSON output. |
check_ai_crawler_access | Fast robots.txt and crawler-access check for every documented AI bot. |
get_fix_plan | Every issue ordered by impact, each with the finding, why it matters, the action and its evidence source. Paste straight into tickets. |
compare_urls | Two to five URLs side by side: your page against competitors, category by category. |
audit_urls | Up to ten URLs scored in one call, worst first. |
export_report | Writes a full markdown or JSON report to a file on your machine. |
How to install
Two ways in. The one-click bundle needs nothing but Claude Desktop. The npx route works in every MCP client and needs Node.js 18 or newer installed.
Claude Desktop: one click
- Download getslurp-mcp.mcpb (720 KB)
- Double-click the file. Claude Desktop opens and asks to install GetSlurp.
- Confirm, then ask Claude: "Audit getslurp.ai and show me the top fixes."
Claude Code: one command
Cursor, Windsurf, Zed and other MCP clients
Add this to your client's MCP configuration (for Claude Desktop the file is claude_desktop_config.json, for Cursor it's mcp.json in settings):
{
"mcpServers": {
"getslurp": {
"command": "npx",
"args": ["-y", "getslurp-mcp"]
}
}
}Restart the client and the six GetSlurp tools appear automatically. Everything runs on your machine: audits go straight from your computer to the sites you check, and no page content, URL or report ever reaches us.
What people ask it to do
- "Audit this article and give me the three fixes with the most impact."
- "Compare our pricing page against these four competitors."
- "Check whether anything in our robots.txt blocks AI search crawlers."
- "Write the fix plan for this URL into a markdown file for the dev team."
The scoring behind every tool is the public GetSlurp methodology.
