Run these tools inside Claude

AutoStaq runs an MCP server at https://autostaq.com/api/mcp. Connect it once and your assistant can audit a page, read a robots.txt or write a content brief without you leaving the chat. No account, no API key, no install.

01

Claude Code

One command in your terminal.

terminal

claude mcp add --transport http autostaq https://autostaq.com/api/mcp

02

Claude desktop and web

Settings, then Connectors, then Add custom connector. Name it AutoStaq and paste the URL below. Leave the OAuth fields empty, the server does not use them.

Server URL

https://autostaq.com/api/mcp

03

Cursor, Windsurf, and anything else

Add this to the MCP config file your client uses.

mcp.json

{
  "mcpServers": {
    "autostaq": {
      "type": "http",
      "url": "https://autostaq.com/api/mcp"
    }
  }
}

04

Then ask it something

Does ChatGPT recommend stripe.com?
Is my pricing page built to be quoted by AI?
Check whether my site blocks AI crawlers
Write me a content brief for “invoicing for freelancers”

The 18 tools it exposes

Tool	What it does
`geo_audit`	We ask ten real buying questions about your category and show you who gets named. Takes a few minutes.
`answer_readiness_score`	Answer-first opening, clean headings, lists, tables, schema, dates, sources. The twelve things that decide whether a model can lift you.
`check_ai_crawlers`	We read your robots.txt and tell you whether GPTBot, ClaudeBot, PerplexityBot and the rest can reach your pages. It is broken more often than people expect.
`generate_llms_txt`	A short file telling AI what your site is and which pages matter. Adoption is still limited, so expect low impact, but it costs nothing.
`extract_entities`	The brands, products, people and topics a model pulls out of your page, and the ones you meant to be there but are not.
`generate_prompt_set`	Every buying question in your category, split by intent, ready to paste into whatever you track with.
`check_robots_txt`	Every rule in your robots.txt, parsed the way crawlers parse it, plus the mistakes that hide your site without telling you.
`check_sitemap`	We find your sitemap, count what is in it, follow the index files, and flag the URLs that should not be there.
`validate_structured_data`	Every JSON-LD block on the page, the types you have, and the required fields you are missing.
`generate_faq_schema`	Paste your questions and answers. Get valid FAQPage JSON-LD, the block models lift most often.
`trace_redirects`	Every hop between the address you type and the page that loads. Chains and loops cost you crawl budget.
`find_broken_links`	We follow every link on the page and report the dead ones, with the status each one returns.
`check_page_speed`	Core Web Vitals from Google, on mobile, with the specific things holding the page back.
`check_heading_structure`	The full outline of your page and every place the hierarchy skips a level. Models read the outline first.
`check_meta_description`	Do you have one, does it fit, and does it earn the click.
`check_readability`	Reading grade, sentence length, passive voice and paragraph size. Dense prose is skipped by readers and skipped by models.
`generate_content_brief`	A full brief for one keyword: angle, outline, questions to answer, entities to name, schema to ship.
`generate_semantic_keywords`	The related terms, entities and questions a page has to cover before it reads as authoritative.

Streamable HTTP transport. Stateless, so there is no session to keep alive.
Rate limited per IP, and the GEO audit shares its budget with the web version.
Every tool works on any public URL. Nothing is stored except the audit cache, for seven days.

More free tools

GEO auditWe ask ten real buying questions about your category and show you who gets named. Takes a few minutes.Prompt gap checkerType any buying question. We ask ChatGPT and show you exactly which brands come back.Answer readiness scoreAnswer-first opening, clean headings, lists, tables, schema, dates, sources. The twelve things that decide whether a model can lift you.AI crawler checkerWe read your robots.txt and tell you whether GPTBot, ClaudeBot, PerplexityBot and the rest can reach your pages. It is broken more often than people expect.llms.txt generatorA short file telling AI what your site is and which pages matter. Adoption is still limited, so expect low impact, but it costs nothing.Entity extractorThe brands, products, people and topics a model pulls out of your page, and the ones you meant to be there but are not.

All free tools

Every other tool tells you you’re losing.

We go fix it. Published, live, done, while you get on with the rest of your week.

Sign up for free to learn more