D

Poor for AI agents

Start here.

These hold your letter down the most
  1. Fail·Security

    Someone who finds a security problem knows whom to tell (security.txt)

    Your site publishes no security.txt, so someone who finds a security problem has no listed way to tell you.

    What to do Publish a security.txt that names whom to contact.

    Your web team · about an hourSee the check

  2. Fail·Act

    AI agents can complete tasks when they visit your site (WebMCP)

    Your home page offers agents no tasks to complete through WebMCP.

    What to do Let agents do every task a customer can do on your site, through WebMCP.

    Your web team · about a daySee the check

  3. Fail·Act

    AI agents can complete tasks without visiting your site (MCP server)

    Your site has no ARD catalog at /.well-known/ard.json, where agents look for an MCP server.

    What to do Give agents a way to do every task a customer can do on your site without visiting it: an MCP server, listed in the site's ARD catalog.

    Your web team · a few daysSee the check

85

Find

Can AI agents and search engines get to your site and list its pages? If they can't find you, you're not in the conversation, and someone else gets recommended.

17% of your grade · 8 of 13 goals pass

  • Your site lists its pages for search engines (XML sitemap)ImportantFail

    Google and Bing use a sitemap to find your pages, though neither promises to list every page in it. A page must be in Google's index to appear in its AI answers. Google uses the dates only while they prove true.

    What we found: We found no valid sitemap at /sitemap.xml or in your robots.txt.

    • Status 404

    How to fix it

    Publish a sitemap of every page you want found, name it in robots.txt, and date each page by when it last changed.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site lists its pages for search engines (XML sitemap) What the grader found: We found no valid sitemap at /sitemap.xml or in your robots.txt. Details: - Status 404 The fix, for a developer: Serve a sitemap.xml (a urlset, or a sitemapindex for several files) of every indexable page at its canonical address, with a lastmod in W3C Datetime from each page's own last edit, and add a Sitemap line with its full address to robots.txt. Source: How to Specify a Canonical with rel="canonical" and Other Methods, Google: https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls; sitemaps.org protocol, XML tag definitions, sitemaps.org: https://www.sitemaps.org/protocol.html#xmlTagDefinitions Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Search results show your site's name and iconImportantFail

    Google shows your site's name above every one of your results, and your icon beside it. Without them it picks a name of its own and shows a blank globe.

    What we found: Your home page has no WebSite structured data, so search engines guess your site's name. Your home page names no icon, so search results show a blank globe beside your name.

    • Your site tells search engines its name
    • Your icon can show beside your results

    How to fix it

    Name the site for search engines and give it a square icon.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Search results show your site's name and icon What the grader found: Your home page has no WebSite structured data, so search engines guess your site's name. Your home page names no icon, so search results show a blank globe beside your name. Details: - Your site tells search engines its name - Your icon can show beside your results The fix, for a developer: Add a WebSite JSON-LD node to the home page with your site's name and its url, and link a square icon of at least 48 by 48 pixels in PNG, ICO or another format Google shows, which robots.txt lets Googlebot-Image fetch. Source: Provide a site name to Google Search, Google: https://developers.google.com/search/docs/appearance/site-names; Define a favicon to show in search results, Google: https://developers.google.com/search/docs/appearance/favicon-in-search Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Every page has a titleImportantPartial

    Google builds the link to your page in its results from the page's title, and may rewrite one that is missing, vague or repeated. Search results show about the first 60 characters. A clear title keeps your own words in front of searchers.

    What we found: 2 of the 12 pages we read share a title, such as "Welcome Home | Zang Apartments | Dallas, TX" on 2 pages.

    • The title "Welcome Home | Zang Apartments | Dallas, TX", which 1 other page has too

    How to fix it

    Give every page its own title that says what the page is and what it offers.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Every page has a title What the grader found: 2 of the 12 pages we read share a title, such as "Welcome Home | Zang Apartments | Dallas, TX" on 2 pages. Details: - The title "Welcome Home | Zang Apartments | Dallas, TX", which 1 other page has too The fix, for a developer: Add a <title> element with text to the head of every page. Describe that page in plain words a person would read, and make it differ from every other page's title. Source: Influencing Title Links in Google Search, Google: https://developers.google.com/search/docs/appearance/title-link; <title>: The Document Title element, MDN: https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/title Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • A page that can be reached at more than one address names its main one (canonical)ImportantPartial

    Google shows only one address for a page. When you name it, the page's standing stays in one place and searchers land on the version you chose, in their language. When the www and the bare address both serve the site, your standing is split between two copies.

    What we found: Your site answers at more than one address: your home page: it moves to your home page with a 302, which says the move is temporary; your home page: it shows the site itself, a second copy; your home page: it moves to your home page with a 302, which says the move is temporary. 1 of the pages we read serves the same text as another address, and no canonical link says which one search engines should show.

    • The same text as your home page, with no canonical link

    How to fix it

    Give each page one main address that search engines treat as its own, and send every other form of your address to it.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: A page that can be reached at more than one address names its main one (canonical) What the grader found: Your site answers at more than one address: your home page: it moves to your home page with a 302, which says the move is temporary; your home page: it shows the site itself, a second copy; your home page: it moves to your home page with a 302, which says the move is temporary. 1 of the pages we read serves the same text as another address, and no canonical link says which one search engines should show. Details: - The same text as your home page, with no canonical link The fix, for a developer: Add one absolute <link rel="canonical"> to each page, pointing at a live, indexable address. Redirect http and the other of www or bare to the site with a 301. Give each hreflang a valid code and a link back, and keep addresses free of fragments, session ids and underscores. Source: SEO Starter Guide: The Basics, Google: https://developers.google.com/search/docs/fundamentals/seo-starter-guide; RFC 6596, The Canonical Link Relation, section 3, IETF: https://www.rfc-editor.org/rfc/rfc6596.html#section-3 Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Your pictures can be found in image searchImportantPartial

    Google's image search finds a picture only in an image element, never one set as a background. The file name is one of the few clues it has about what a picture shows.

    What we found: Image files are named in words (1 of 3 image files).

    • “Z 01.JPG” names nothing that is in the picture: <img src="images/specific-graphics/photo-gallery/Z 01.JPG" alt="Zang Apartments community" class="hero-image" fetchpriority="high">

    How to fix it

    Show important pictures as images and name each file in words.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your pictures can be found in image search What the grader found: Image files are named in words (1 of 3 image files). Details: - “Z 01.JPG” names nothing that is in the picture: <img src="images/specific-graphics/photo-gallery/Z 01.JPG" alt="Zang Apartments community" class="hero-image" fetchpriority="high"> The fix, for a developer: Use an <img> element for every picture that carries meaning, keep CSS backgrounds for decoration, and name each file for what it shows, such as blue-sofa.jpg instead of IMG_0023.jpg. Source: Image SEO Best Practices, Google: https://developers.google.com/search/docs/appearance/google-images Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your site's security doesn't turn AI agents away (firewall)CriticalPass

    Many sites sit behind a firewall or bot-protection service that blocks automated visits. An agent stopped at the door reads nothing, so it answers from other sites. Claude's web fetch, for example, gets only what your site sends back.

  • Your site lets AI agents in (robots.txt)CriticalPass

    AI services send crawlers, robots that read sites for them. One your robots.txt blocks cannot use your pages, so answers about your business come from other sites. OpenAI, Anthropic and Perplexity each say their crawlers obey robots.txt.

  • Search engines may show your pages in their resultsCriticalPass

    A page search engines may not list is missing from Google's and Bing's results, and from AI answers built on them, such as Google's AI Overviews. Brave's search, which some AI agents use, follows Google's rules.

  • Your instructions to AI agents and search engines have no errors (robots.txt)ImportantPass

    Google's crawler skips any line of robots.txt it cannot read, so a rule you meant to set never applies. A page blocked in one place and allowed in another is read the wrong way. A page that forbids quoting cannot appear in Google's AI Overviews.

  • Your site says when a page does not exist (404)ImportantPass

    When a missing page says so, Google drops the address from search. When it answers like a normal page, Google flags a soft 404, and every mistyped address looks like a real page.

  • Your pages open without unnecessary redirectsImportantPass

    Each extra redirect is another trip across the network that can delay the page by hundreds of milliseconds. Every visitor waits for every hop.

  • AI agents are sent the same page people areImportantPass

    A site that shows agents something different, or a block, is answered about wrongly or not at all. Google's spam policies call showing robots different content cloaking. Claude's web fetch, for example, reads only what your site sends it.

38

Read

Can agents quickly work out who you are and what you offer? An agent only spends so much effort on one site. If yours is hard to read, it moves on and recommends one that's easier.

17% of your grade · 4 of 12 goals pass

  • Your site tells AI agents what your business is (structured data)CriticalFail

    Structured data saves agents and search engines from guessing your type of business and contact details from the page. Google says it gives explicit clues about what a page means. Bing says it may help and promises nothing.

    What we found: Your home page has no structured data with a type.

    • Your home page has no structured data with a type

    How to fix it

    Describe the business in structured data, using the most specific schema.org type that fits it.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site tells AI agents what your business is (structured data) What the grader found: Your home page has no structured data with a type. Details: - Your home page has no structured data with a type The fix, for a developer: Add a JSON-LD script to the home page whose "@type" is the most specific schema.org type that fits the business. schema.org lists every type and its subtypes. Source: Intro to structured data, Google: https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data; JSON-LD 1.1, W3C: https://www.w3.org/TR/json-ld11/ Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your site has a short guide for AI agents (llms.txt)CriticalFail

    llms.txt gives agents a short map of your site, so they reach your key pages without reading every one. Google's Lighthouse checks it as an emerging convention, and the grade counts standards likely to be adopted.

    What we found: We found no /llms.txt on your site.

    • There is no /llms.txt (it answered 404).

    How to fix it

    Publish /llms.txt with your business name as its title and links to your most important pages.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site has a short guide for AI agents (llms.txt) What the grader found: We found no /llms.txt on your site. Details: - There is no /llms.txt (it answered 404). The fix, for a developer: Serve a Markdown file at /llms.txt with a 200, and keep it clear of any rule that sends unknown paths to the home page or a firewall page. Start it with an H1 naming the site, then sections of links to key pages that answer 200, as llmstxt.org describes. Source: llms.txt audit, Chrome: https://developer.chrome.com/docs/lighthouse/agentic-browsing/llms-txt; The /llms.txt file, format, llmstxt.org: https://llmstxt.org/#format Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Your site tells AI agents everything they can do with it, in one place (ARD catalog)CriticalFail

    Agents read the catalog before they open a page, so they know straight away which of your services they can use, such as your MCP server. Google's Lighthouse checks it too, and the grade counts standards likely to be adopted, such as this draft.

    What we found: We found no ARD catalog at /.well-known/ard.json on your site.

    • There is no ARD catalog (it answered 404).

    How to fix it

    Publish an ARD catalog that lists every tool and service agents can use, such as your MCP server.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site tells AI agents everything they can do with it, in one place (ARD catalog) What the grader found: We found no ARD catalog at /.well-known/ard.json on your site. Details: - There is no ARD catalog (it answered 404). The fix, for a developer: Serve JSON at /.well-known/ard.json with an entries list that follows the ARD schema: each entry has a urn:air: identifier, a displayName, a type and a url. An MCP server's entry has the type application/mcp-server-card+json and points at its server card, at /mcp/server-card. Link it from every page with <link rel="ard" href="/.well-known/ard.json">. A catalog found only at the older /.well-known/ai-catalog.json does not count. Source: Lighthouse agentic browsing scoring, Chrome: https://developer.chrome.com/docs/lighthouse/agentic-browsing/scoring; RFC 9727: api-catalog, a Well-Known URI and Link Relation to Help Discovery of APIs, IETF: https://www.rfc-editor.org/rfc/rfc9727.html Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your site sends AI agents a plain-text version of each page (Markdown)CriticalFail

    Agents read Markdown faster and more accurately than a full web page, and it costs them fewer tokens, the units AI services count text in. Cloudflare calls Markdown the common language of agents.

    What we found: We asked your home page for Markdown and got none back.

    • Answered with text/html

    How to fix it

    Send each page as Markdown to agents that ask for it.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site sends AI agents a plain-text version of each page (Markdown) What the grader found: We asked your home page for Markdown and got none back. Details: - Answered with text/html The fix, for a developer: Answer a request whose Accept header names text/markdown with the page's content as Markdown and Content-Type: text/markdown, and link it from the page with <link rel="alternate" type="text/markdown">. Cloudflare can convert pages at its edge, or the server can render the Markdown itself. Source: Markdown for Agents, Cloudflare: https://developers.cloudflare.com/fundamentals/reference/markdown-for-agents/; Link Relations: alternate, IANA: https://www.iana.org/assignments/link-relations/link-relations.xhtml Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your structured data is correct and completeImportantFail

    One stray comma and a machine throws the whole block away, and a property it does not know is ignored. Google recommends JSON-LD, may drop rich results for data that claims what the page does not show, and forbids star ratings a business gives itself.

    What we found: Your home page's structured data does not say what business you are: no node in it is an Organization or a LocalBusiness.

    • Your home page says, as data, what business you are

    How to fix it

    Fix the mistakes this report lists in the site's structured data.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your structured data is correct and complete What the grader found: Your home page's structured data does not say what business you are: no node in it is an Organization or a LocalBusiness. Details: - Your home page says, as data, what business you are The fix, for a developer: Describe the business in one JSON-LD node with an @id every page points to, with logo, telephone, address, hours and sameAs. Use valid JSON and only schema.org's types and properties, ISO 8601 dates, hours as Mo-Fr 09:00-17:00, 5-decimal positions, each rich result's required properties, breadcrumbs, and only what the page shows. Source: Intro to structured data, Google: https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data; JSON-LD 1.1, section 3.3 Node Identifiers, W3C: https://www.w3.org/TR/json-ld11/#node-identifiers Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your site says who runs it and who wrote whatImportantFail

    People and search engines both ask who is behind a claim and how fresh it is. Google's guide to helpful content asks whether a page makes clear who wrote it and who runs the site. A visitor who cannot tell who you are does not get in touch.

    What we found: The home page doesn't say who runs the site or how to reach them.

    • Nothing says who runs the site: no link to an about page, no address element, and no structured data naming the business with its address.

    How to fix it

    Link an about page and a way to reach you from your home page, and put the author's name and the date on every article.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site says who runs it and who wrote what What the grader found: The home page doesn't say who runs the site or how to reach them. Details: - Nothing says who runs the site: no link to an about page, no address element, and no structured data naming the business with its address. The fix, for a developer: Link the about and contact pages, or tel: and mailto: links, from the home page's header or footer, and give each Article an author with a name and a datePublished in its JSON-LD and on the page. Source: Creating helpful, reliable, people-first content, Google: https://developers.google.com/search/docs/fundamentals/creating-helpful-content Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    You · about an hour
  • Every page has a summary for search resultsImportantPartial

    Google may show your summary under the page's title in its results, and asks for a different one on every page. Without one, it picks a snippet from the page, which may be a line you would never choose. A shared link with no picture or title is easy to pass over.

    What we found: 12 of the 12 pages we read share a description, such as "Zang Apartments in Dallas, Texas. One and two bedroom apartments. Call (214) 941-7444." on 12 pages. Your home page has no og:title, og:type, og:image, og:url, so a shared link shows no proper card.

    • The description "Zang Apartments in Dallas, Texas. One and two bedroom apartments. Call (214) 941-7444.", which 11 other pages have too
    • No og:title

    How to fix it

    Write a one or two sentence summary of every page, different on each.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Every page has a summary for search results What the grader found: 12 of the 12 pages we read share a description, such as "Zang Apartments in Dallas, Texas. One and two bedroom apartments. Call (214) 941-7444." on 12 pages. Your home page has no og:title, og:type, og:image, og:url, so a shared link shows no proper card. Details: - The description "Zang Apartments in Dallas, Texas. One and two bedroom apartments. Call (214) 941-7444.", which 11 other pages have too - No og:title The fix, for a developer: Add <meta name="description" content="..."> to the head of every page, with a sentence or two unique to that page, and og:title, og:type, og:url and an og:image that loads. Source: Document does not have a meta description, Chrome: https://developer.chrome.com/docs/lighthouse/seo/meta-description; How to Write Meta Descriptions, Google: https://developers.google.com/search/docs/appearance/snippet Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Each page marks its header, main content and footer (landmarks)ImportantPartial

    Agents and screen readers jump between a page's regions to reach the part they need, such as the main content past the navigation. Content outside every region, or two navigation regions with the same name, leaves them reading the page from the top.

    What we found: Some content sits outside every marked region of the page.

    • Its content sits outside every landmark: <section class="closing">

    How to fix it

    Mark each page's header, main content and footer as regions, and put everything on the page inside one.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Each page marks its header, main content and footer (landmarks) What the grader found: Some content sits outside every marked region of the page. Details: - Its content sits outside every landmark: <section class="closing"> The fix, for a developer: Use one header, one main and one footer element at the top level of the page, keep every piece of content inside a landmark, and give landmarks of the same kind, such as two nav elements, different aria-labels. Source: Learn Accessibility: Content structure, Google: https://web.dev/learn/accessibility/structure; Understanding Success Criterion 1.3.1: Info and Relationships, W3C: https://www.w3.org/WAI/WCAG22/Understanding/info-and-relationships.html Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • AI agents can read your text without running scriptsCriticalPass

    Many agents read only the page your server sends and run none of its scripts. If your text appears only after scripts run, those agents see an empty page and answer from other sites. Claude's web fetch is one of them.

  • AI agents get your words, links and data without running scriptsCriticalPass

    Many agents read only the page your server sends and run none of its scripts, so anything a script draws, structured data included, is invisible to them. Claude's web fetch is one of them.

  • Your pages are clear to readImportantPass

    One main heading gives people and machines the same starting point. Text inside a picture cannot be read aloud, searched or quoted, and WCAG, the accessibility standard most laws point to, asks for text where text will do.

  • Nothing covers your page when a visitor arrives on a phoneImportantPass

    A pop-up over the first screen, or a bar that takes more than 30% of a phone screen, is among the experiences people most dislike, the Coalition for Better Ads found. Some agents read the page as a picture and see only what is on top.

29

Act

Can agents do what a customer asks, such as book, reserve, enquire or search? When someone asks an agent to act for them and your site doesn't let it, the agent acts on a site that does.

17% of your grade · 1 of 5 goals pass

  • AI agents can complete tasks when they visit your site (WebMCP)CriticalFail

    Your site can tell agents what it lets a customer do, for example book a tour, reserve a table, ask a question, or search what's available. An agent can then do it for them, start to finish. Without that, the agent can only read your pages and send the customer to do it themselves, or do it on another site that lets it.

    How to fix it

    Let agents do every task a customer can do on your site, through WebMCP.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: AI agents can complete tasks when they visit your site (WebMCP) What the grader found: Your home page offers agents no tasks to complete through WebMCP. The fix, for a developer: Register a WebMCP tool for every task a customer can do as the page loads: call document.modelContext.registerTool() with a name, a description, an input schema and an execute function, or add toolname and tooldescription attributes to the form. A name is 1 to 128 letters, digits, _, - or . Source: WebMCP for developers, Chrome: https://developer.chrome.com/docs/ai/webmcp; WebMCP, Draft Community Group Report, W3C Web Machine Learning Community Group: https://webmachinelearning.github.io/webmcp/ Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • AI agents can complete tasks without visiting your site (MCP server)CriticalFail

    Some agents never open a web page. They connect to a business directly and work from there. Giving them a way in means your site can be searched and booked by those agents too.

    What we found: Your site has no ARD catalog at /.well-known/ard.json, where agents look for an MCP server.

    • /.well-known/ard.json answers 404

    How to fix it

    Give agents a way to do every task a customer can do on your site without visiting it: an MCP server, listed in the site's ARD catalog.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: AI agents can complete tasks without visiting your site (MCP server) What the grader found: Your site has no ARD catalog at /.well-known/ard.json, where agents look for an MCP server. Details: - /.well-known/ard.json answers 404 The fix, for a developer: Serve the MCP server card at /mcp/server-card, as the server card extension places it, and list it in /.well-known/ard.json as an entry with a urn:air: identifier, a displayName, the type application/mcp-server-card+json and the card's URL. Source: SEP-2127: MCP Server Cards, Model Context Protocol: https://modelcontextprotocol.io/seps/2127-mcp-server-cards; Agentic Resource Discovery specification, agenticresourcediscovery.org: https://agenticresourcediscovery.org/spec/ Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · a few days
  • AI agents and visitors can fill in and send your formsCriticalFail

    Fields that say they ask for an email or a phone fill in correctly, for browsers and agents alike. A see-through layer over a button stops a click from landing, a puzzle turns away agents and many people with disabilities, and a form that needs a script loses the enquiry when the script fails.

    What we found: A button or link sits under an invisible layer that catches the click.

    • An invisible element sits over it at its centre: <div class="launcher__mobile-launcher-wrapper" style="bottom:72px;right:18px;">: <a class="hero-scroll" href="#welcome">

    How to fix it

    Make every enquiry form one an agent or a visitor can fill in and send, with no puzzle in the way.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: AI agents and visitors can fill in and send your forms What the grader found: A button or link sits under an invisible layer that catches the click. Details: - An invisible element sits over it at its centre: <div class="launcher__mobile-launcher-wrapper" style="bottom:72px;right:18px;">: <a class="hero-scroll" href="#welcome"> The fix, for a developer: Give each field its autocomplete token and type (email, tel, name, street-address), put fields in a <form> with a submit button and an action, remove invisible layers over controls, and replace a visible CAPTCHA with an invisible check. Source: Build agent-friendly websites, web.dev: https://web.dev/articles/ai-agent-site-ux; HTML Standard, Autofill, WHATWG: https://html.spec.whatwg.org/multipage/form-control-infrastructure.html#autofill Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Visitors on a phone can reach your main actionImportantPartial

    Most viewing time goes to what shows before any scrolling, Nielsen Norman Group found. People and agents learn where to act once and expect it there again, and a navigation button that does not open leaves every page behind it out of reach.

    What we found: Neither your home page nor the 11 pages it links that we read holds a form or a booking widget a shopper can use to reach you.

    • No contact form or booking widget on the home page or the pages it links

    How to fix it

    Put your main action on the first screen and in the header of every page, and link a contact or booking form from the home page.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Visitors on a phone can reach your main action What the grader found: Neither your home page nor the 11 pages it links that we read holds a form or a booking widget a shopper can use to reach you. Details: - No contact form or booking widget on the home page or the pages it links The fix, for a developer: Place the primary call to action above the fold at 412 px wide, keep it in the shared header, make the header's navigation button open its links on tap, and link the contact or booking form from the home page. Source: 'Contact Us' Page Guidelines, Nielsen Norman Group: https://www.nngroup.com/articles/contact-us-pages/; Understanding Success Criterion 3.2.3: Consistent Navigation, W3C: https://www.w3.org/WAI/WCAG22/Understanding/consistent-navigation.html Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • AI agents can see your site's buttons and fields (accessibility tree)CriticalPass

    Agents find what they can click or fill in through the accessibility tree, the browser's list of a page's controls. A control missing from it does not exist for an agent, so it cannot book, buy or send an enquiry. Screen readers use the same tree.

89

Performance

Does your site answer fast? Agents work against the clock, and so do people. A slow site costs them time, and they give that time to a faster one.

Speed comes from real visitors of your page over the last 28 days when Google has their numbers, and from a test on a phone when it does not.

17% of your grade · 5 of 9 goals pass

  • Your main content shows up quickly on phonesCriticalPartial

    Visitors leave pages that are slow to show anything, and Google uses this speed in ranking. It is called Largest Contentful Paint, one of Google's three Core Web Vitals.

    What we found: The main content takes 5.9 s to show. Lab test on a phone.

    • Holds the first paint 224 ms
    • In the longest chain, 353 ms from the first request

    How to fix it

    Make the first screen load faster on phones, starting with its biggest image.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your main content shows up quickly on phones What the grader found: The main content takes 5.9 s to show. Lab test on a phone. Details: - Holds the first paint 224 ms - In the longest chain, 353 ms from the first request The fix, for a developer: Serve the largest element early: preload it, avoid render-blocking scripts and styles, and cut server response time. The report lists what holds it back. Source: Web Vitals, Google: https://web.dev/articles/vitals; Largest Contentful Paint (LCP), Google: https://web.dev/articles/lcp Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · a week or more
  • Images are sent at the size they are shownImportantPartial

    An image bigger than the space it fills wastes data and slows the page, most of all on a phone connection. Smaller images mean a faster page for every visitor.

    What we found: Improve image delivery.

    • Could be 501 KiB smaller: a modern format or more compression: <img src="images/specific-graphics/photo-gallery/Z 01.JPG" alt="Zang Apartments community" class="hero-image" fetchpriority="high">

    How to fix it

    Resize and compress your images before they go on the page.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Images are sent at the size they are shown What the grader found: Improve image delivery. Details: - Could be 501 KiB smaller: a modern format or more compression: <img src="images/specific-graphics/photo-gallery/Z 01.JPG" alt="Zang Apartments community" class="hero-image" fetchpriority="high"> The fix, for a developer: Serve several sizes with srcset and sizes, use a modern format such as WebP or AVIF, and keep each image's width and height in the proportion it is shown. Source: Properly size images, Chrome: https://developer.chrome.com/docs/lighthouse/performance/uses-responsive-images Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Your pages are light to downloadImportantPartial

    Heavy pages take longer to load and cost visitors on mobile data real money. Chrome says large downloads are closely tied to long load times, and uncompressed text is several times larger than it needs to be.

    What we found: Reduce unused JavaScript (137 KiB unused).

    • 137 KiB of 231 KiB unused, from eliseai.com

    How to fix it

    Remove code and files the page does not use.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your pages are light to download What the grader found: Reduce unused JavaScript (137 KiB unused). Details: - 137 KiB of 231 KiB unused, from eliseai.com The fix, for a developer: Minify scripts and styles, remove unused and duplicate code, stop shipping legacy polyfills, compress every text response with Brotli or gzip, and serve fonts as WOFF2. The report lists the files by size. Source: Avoid enormous network payloads, Chrome: https://developer.chrome.com/docs/lighthouse/performance/total-byte-weight; Compression in HTTP, MDN: https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/Compression Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · a few days
  • Return visits load fasterImportantPartial

    Downloading every file again on each visit is slow and costly. When browsers keep the files, returning visitors see your page sooner, and a page kept for days shows old prices and hours. People go back and forth between pages while they compare.

    What we found: Use efficient cache lifetimes.

    • Not cached

    How to fix it

    Let browsers keep images, scripts and styles between visits, check pages for changes, and let the Back button restore them.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Return visits load faster What the grader found: Use efficient cache lifetimes. Details: - Not cached The fix, for a developer: Send Cache-Control with a long max-age for fingerprinted static files, no-cache or a max-age of ten minutes or less for HTML, and remove what keeps the page from the back/forward cache, such as unload handlers and Cache-Control: no-store. Source: Prevent unnecessary network requests with the HTTP Cache, Google: https://web.dev/articles/http-cache; HTTP caching, MDN: https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/Caching Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Your site responds quickly to taps and clicksCriticalPass

    A page that freezes after a tap feels broken, so visitors tap again or leave. Google rates a reaction under 200 milliseconds as good and counts it among its Core Web Vitals.

  • Your pages stay still while they loadCriticalPass

    When content jumps as a page loads, visitors tap the wrong thing, such as a link that moved under their finger. Google counts layout shift among its three Core Web Vitals.

  • Your pages load without errorsImportantPass

    An error in the browser often means part of your page is broken, such as a form that will not send. Chrome says each one is a problem to resolve.

  • Your site does not ask visitors for permissions as they arriveImportantPass

    People distrust or are confused by a page that asks for their location the moment it opens. A field that blocks pasting makes a form harder to fill in.

  • Your server answers quicklyImportantPass

    Every visitor and agent waits for your server's first response before anything else can load. Chrome says a CDN, a network of servers close to your visitors, cuts that wait.

55

Security

Is what your site serves safe for the people who visit? Browsers warn people away from a page that isn't secure, and an agent shouldn't send a shopper's details to one.

17% of your grade · 1 of 4 goals pass

  • Someone who finds a security problem knows whom to tell (security.txt)ImportantFail

    Without a contact, a person who finds a hole in your site has no safe way to report it, so it may never be reported. RFC 9116 defines the file for exactly that.

    What we found: Your site publishes no security.txt, so someone who finds a security problem has no listed way to tell you.

    • No security.txt at /.well-known/security.txt or /security.txt

    How to fix it

    Publish a security.txt that names whom to contact.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Someone who finds a security problem knows whom to tell (security.txt) What the grader found: Your site publishes no security.txt, so someone who finds a security problem has no listed way to tell you. Details: - No security.txt at /.well-known/security.txt or /security.txt The fix, for a developer: Serve /.well-known/security.txt as text/plain with a Contact line (a mailto: or a form's address) and an Expires date under a year away, as RFC 9116 defines it. Source: RFC 9116: A File Format to Aid in Security Vulnerability Disclosure, IETF: https://www.rfc-editor.org/rfc/rfc9116; security.txt, securitytxt.org: https://securitytxt.org/ Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Your site loads over a secure connectionImportantPartial

    A secure connection stops others reading or changing what passes between your site and its visitors. Browsers mark pages without one as not secure, and Google counts it in page experience. An expired certificate replaces your site with a browser warning.

    What we found: Your page sends no Strict-Transport-Security header, so browsers are never told to keep to its secure address.

    • The page sends no Strict-Transport-Security header.

    How to fix it

    Serve every page over a secure connection and send the plain address to it.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your site loads over a secure connection What the grader found: Your page sends no Strict-Transport-Security header, so browsers are never told to keep to its secure address. Details: - The page sends no Strict-Transport-Security header. The fix, for a developer: Serve every page and file over a valid certificate that is logged and renewed well before it ends, redirect http to https in one hop, send Strict-Transport-Security with a max-age of at least six months, refuse the protocol versions before 1.2, and send no form or resource over plain http. Source: Why HTTPS matters, Google: https://web.dev/articles/why-https-matters; Strict-Transport-Security, MDN: https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Headers/Strict-Transport-Security Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about an hour
  • Your pages tell browsers how to keep your visitors safeImportantPartial

    These rules limit the harm when someone slips code into a page, or tricks a visitor into clicking a hidden copy of it. A stolen cookie can let someone act as your visitor. OWASP, the web security project, recommends them for every site.

    What we found: Your page sends no Content-Security-Policy, so a browser runs any script that ends up in it. Your page does not tell browsers to keep other sites from showing it in a frame. Your page does not send X-Content-Type-Options: nosniff, so browsers may guess what a file is.

    • No Content-Security-Policy header or meta tag.
    • No frame-ancestors in a Content-Security-Policy header and no X-Frame-Options header.
    • No X-Content-Type-Options header.

    How to fix it

    Add the security headers this report lists.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your pages tell browsers how to keep your visitors safe What the grader found: Your page sends no Content-Security-Policy, so a browser runs any script that ends up in it. Your page does not tell browsers to keep other sites from showing it in a frame. Your page does not send X-Content-Type-Options: nosniff, so browsers may guess what a file is. Details: - No Content-Security-Policy header or meta tag. - No frame-ancestors in a Content-Security-Policy header and no X-Frame-Options header. - No X-Content-Type-Options header. The fix, for a developer: Send a Content-Security-Policy whose script-src uses nonces or hashes and sets frame-ancestors, X-Content-Type-Options: nosniff and a Referrer-Policy of strict-origin-when-cross-origin; never pair a reflected Origin with credentials; mark cookies Secure, and session cookies HttpOnly. Source: HTTP Headers Cheat Sheet, OWASP: https://cheatsheetseries.owasp.org/cheatsheets/HTTP_Headers_Cheat_Sheet.html; Cross-Origin Resource Sharing (CORS), MDN: https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/CORS Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • The code your pages load is safeImportantPass

    Attackers scan for exactly the versions with published holes, and out-of-date site software is a common way in, OWASP says. A key left in public code is found and used, and card-skimming code is slipped in through payment pages.

94

Compliance

Can everyone use your site, including people who rely on a screen reader or a keyboard, and does it follow the rules that apply to it? Agents read your controls the way screen readers do, so a site that shuts people out shuts agents out too.

17% of your grade · 10 of 12 goals pass

  • Your text is clear to read on any screenImportantPartial

    Some agents look at a page as a picture, as people do, and pale text is easy for both to miss. The same contrast lets people with low vision read it. WCAG, the accessibility standard most laws point to, sets how much contrast text needs.

    What we found: Some text is too pale against its background. 29 more elements' text needs a look: it sits over an image, a gradient or another layer, or is one character or symbols alone.

    • Text at 4.43:1 contrast (#646b66 on #e8e8de), under the 4.5:1 it needs: <p>
    • Text at 4.42:1 contrast (#925d38 on #e8e8de), under the 4.5:1 it needs: <span>

    How to fix it

    Darken pale text, underline links inside paragraphs, let visitors zoom, and fix the text this report finds cut off or overlapping.

    Fix with AI

    My website https://zangapts.com/ fell short on one check of Clutch's Agent Grader, which tests whether AI agents can find, read and use a site. Check: Your text is clear to read on any screen What the grader found: Some text is too pale against its background. 29 more elements' text needs a look: it sits over an image, a gradient or another layer, or is one character or symbols alone. Details: - Text at 4.43:1 contrast (#646b66 on #e8e8de), under the 4.5:1 it needs: <p> - Text at 4.42:1 contrast (#925d38 on #e8e8de), under the 4.5:1 it needs: <span> The fix, for a developer: Reach a contrast ratio of 4.5 to 1 for normal text, mark links in text with more than color, allow zoom, leave !important off inline spacing, remove blink and marquee, let boxes grow with their text, keep content within the screen's width, and stop looping motion under prefers-reduced-motion. Source: Learn Accessibility: Color and contrast, Google: https://web.dev/learn/accessibility/color-contrast; Understanding Success Criterion 1.4.3: Contrast (Minimum), W3C: https://www.w3.org/WAI/WCAG22/Understanding/contrast-minimum.html Explain the fix in plain words, then write what my web team should change and how we can check it worked. Full report, if you want more context: https://clutch.io/grade/zangapts-com/2026-11-20

    Your web team · about a day
  • Headings give each page a clear outlineImportantPass

    Clear headings show search engines and agents how a page is laid out, so they find the part that answers a question, such as your opening hours. Google's guide to its AI features recommends them.

  • AI agents can tell what your images show (alt text)ImportantPass

    Agents rely on an image's text description, called alt text, to know what it shows, the same description that lets people who cannot see the image use the page. Google calls alt text the most important attribute for describing an image.

  • Your pages say which language they are inImportantPass

    A declared language tells software how to read your page, from the voice a screen reader uses to the language a translation starts from, so the wrong one comes out as noise. WCAG, the accessibility standard most laws point to, requires it.

  • Your lists and tables are built so agents can read themImportantPass

    Agents read a list or a table as one only when the page's code marks it as one. Otherwise your prices or opening hours reach them as loose lines. The same code lets people who use screen readers follow them, and WCAG requires it.

  • AI agents can tell what custom controls are, such as date pickersImportantPass

    A custom date picker or slider is invisible to an agent unless its code gives it a name, so the agent cannot choose a date or set a price range. The same name tells people who use screen readers what the control is.

  • Hidden labels on your buttons and fields are accurate (ARIA)ImportantPass

    Agents trust these labels. A wrong one tells an agent a control is something it is not, such as a button marked as a heading, so it clicks the wrong thing or nothing. Screen readers trust them too, so people with disabilities are misled the same way.

  • Buttons and links are big enough for an agent to clickImportantPass

    Agents that click where a button sits, as a person does, miss small or crowded targets, and so do people on phones and people with limited movement. Google's web.dev recommends targets of about 48 pixels, and WCAG sets a minimum of 24.

  • AI agents can move around your pages without a mouseImportantPass

    Agents move through a page from control to control, as keyboard users do. A page that skips controls or refreshes on its own keeps them from the button they need. WCAG requires every function to work from a keyboard, which people with disabilities rely on.

  • Your site respects visitors' privacyImportantPass

    California's law asks every commercial site to link its privacy policy from the home page. A study of 100,000 sites found trackers taking email addresses as they were typed, and the FTC calls a choice that changes nothing a design that subverts privacy.

  • Your site deals fairly with visitorsImportantPass

    The FTC names pre-ticked boxes and countdowns that are not really timed among the dark patterns it acts on, and a claim you cannot back is what a complaint is built on. An accessibility statement gives a person who meets a barrier somewhere to turn.

Fix it,

or see it rebuilt

These 26 fixes add up to about five weeks of your web team's time.

  • Send to your agency

    An email with what is wrong and a link to this report, so whoever runs the site has every fix and its line for the web team.

    Send to your agency
  • Explore with your AI

    Your own AI reads the report and tells you how to fix each issue and the effort each takes. It gets this, and nothing else:

    Read the Agent Grader report for zangapts.com, which tests whether AI agents can find, read and use the site: https://clutch.io/grade/zangapts-com/2026-11-20 Then tell me how to fix each issue it lists, in order, and the effort each one takes, in plain words I can pass to my web team.

    Fix with AI

    Read the Agent Grader report for zangapts.com, which tests whether AI agents can find, read and use the site: https://clutch.io/grade/zangapts-com/2026-11-20 Then tell me how to fix each issue it lists, in order, and the effort each one takes, in plain words I can pass to my web team.

  • See your site rebuilt

    Clutch rebuilds zangapts.com from its address, and the new site doesn't have these issues. Free to preview, $39 a month once it goes live.

    See your site rebuilt

See if AI chooses you

This scan graded the site on its own. Grade it again and an AI agent shops for it too, with a search your site already ranks for, and shows where it got through and where it stopped.

See zangapts.com rebuilt, without these issues

2nd of 3 residence sites graded so far. See the board.

How we grade