mastria.com — AI Site Grade

Analysis

The site is entirely invisible to AI crawlers

Mastria Auto Group operates a live dealership website that returns HTTP 403 Forbidden to every single request — from browsers, from all major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, OAI-SearchBot, Bytespider, ChatGPT-User), and even to its own robots.txt and llms.txt files. The site is behind Cloudflare with no bypass mechanism. The Wayback Machine snapshot from March 2025 shows a fully functional site with inventory, service pages, and financing tools — but the live site is a JS shell that serves zero visible content to any automated agent.

Crawler Access

Every bot tested receives a 403 response from Cloudflare. The robots.txt at https://mastria.com/robots.txt returns a 403 with a 236KB HTML payload (a Cloudflare challenge page), not a text file. The llms.txt at https://mastria.com/llms.txt returns the same 403 HTML shell. No AI crawler can discover any page, any schema, any content, or any crawl directive. The site has no robots.txt rules because the file is inaccessible. The DNS resolves to 74.119.99.3 with SPF records referencing _spf.hosting.cdkglobal.com — confirming this is a CDK Global dealer platform deployment, which typically uses heavy JavaScript rendering and Cloudflare protection.

Cold-Knowledge Gap

LLM knowledge about "Mastria" is zero. The model reports no verifiable information about the brand, cannot confirm what it does, who it serves, or any products. This is a complete knowledge vacuum. The Wayback Machine reveals the site is Mastria Auto Group, a multi-brand dealership in Raynham, MA (1305 New State Highway, Raynham, MA 02767) selling Cadillac, GMC, Kia, Mazda, Nissan, Subaru, and Volkswagen. The site had rich content: a "7/30 Worry-Free Guarantee," online buying (FastLane), home delivery, towing services, and a full service/collision center. None of this is accessible to AI models.

Schema Posture

The archived homepage contained a well-structured AutoDealer schema with PostalAddress, OpeningHoursSpecification, and three departments (AutoPartsStore, AutoDealer for Sales, AutoRepair). The schema included telephone numbers, price ranges, and images. This schema is completely invisible to AI crawlers on the live site. The live site returns no JSON-LD, no meta description, no canonical, and no heading structure — just a generic <title>Dealer Website</title> and a Cloudflare wall.

External Signals

External search engines return zero indexed results for "Mastria Auto Group," "Mastria Raynham," or any variant. The brand has no discoverable Reddit threads, press mentions, or review aggregator pages in search results. The Wayback Machine shows DealerRater reviews embedded on the homepage (dated March 2025, mentioning sales staff Ed Cappola, Derek Welch, and service advisor Dylan), but these reviews are not surfaced anywhere AI models can reach. The site's SPF record includes Google Workspace, suggesting active email — but the web presence is a ghost.

Findings

1. All AI crawlers receive HTTP 403 from Cloudflare High

   Every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, OAI-SearchBot, Bytespider, ChatGPT-User) receives a 403 Forbidden response from Cloudflare. No bot can access any page, including robots.txt and llms.txt.

   What to change: Configure Cloudflare to allow AI crawler user agents by creating a WAF bypass rule or serving a static robots.txt that permits GPTBot, ClaudeBot, and other listed bots.

2. robots.txt returns 403 HTML shell instead of text file High

   The robots.txt at https://mastria.com/robots.txt returns a 403 status with a 236KB Cloudflare challenge page, not a plain-text robots file. No crawl directives are available to any bot.

   What to change: Serve a valid robots.txt file from the origin that allows AI crawlers to access the site.

3. llms.txt returns 403 and is not published Medium

   The llms.txt file at https://mastria.com/llms.txt returns a 403 Forbidden response, indicating it is either not published or blocked by Cloudflare. AI crawlers cannot discover any structured content guidance.

   What to change: Publish an llms.txt file that lists key pages (inventory, service, about) and any available LLM-friendly content.

4. Live site serves empty JS shell with no visible content High

   The live homepage returns a generic '<title>Dealer Website</title>' and a Cloudflare challenge page with zero visible text content. No inventory, service information, or brand messaging is accessible to any automated agent.

   What to change: Implement server-side rendering (SSR) or static prerendering so that AI crawlers receive HTML content without executing JavaScript.

5. Zero indexed pages in search engines High

   Multiple web searches for 'Mastria Auto Group', 'Mastria Raynham', and 'mastria.com' return zero results. The site has no organic search presence, making it invisible to both users and AI models that rely on search indices.

   What to change: Ensure the site is indexable by search engines by removing the blanket 403 block and implementing proper SEO fundamentals (meta tags, sitemap, server-side content).

6. LLMs have zero verifiable knowledge about Mastria High

   The LLM knowledge base contains no information about Mastria Auto Group, its brands, location, or services. The brand is completely absent from AI training data.

   What to change: Publish authoritative content about the dealership on the live site and ensure it is accessible to AI crawlers so that LLMs can index it.

7. AutoDealer schema exists in archive but is invisible on live site High

   The archived homepage contained a well-structured AutoDealer schema with address, hours, departments, and contact info. The live site returns no JSON-LD or structured data, making all schema invisible to AI crawlers.

   What to change: Reinstate the AutoDealer JSON-LD schema on the live site and ensure it is served in the initial HTML (not injected via JavaScript).

8. No external mentions or reviews found in search results Medium

   Searches for Mastria Auto Group on Reddit, review sites, and news return zero results. The brand has no discoverable external signals that AI models could use to verify its existence or reputation.

   What to change: Encourage customers to leave reviews on major platforms (Google, DealerRater) and engage in local SEO to build external citations.

9. Cloudflare challenge page blocks all automated access High

   The site uses Cloudflare's JavaScript challenge, which returns a 403 with a challenge page to any request that does not execute JavaScript. This blocks all AI crawlers and even simple HTTP clients.

   What to change: Configure Cloudflare to allow known AI crawler user agents by creating a WAF rule that bypasses the challenge for those bots.

10. No sitemap discovered or accessible Medium

    The tool could not discover any sitemap URLs for mastria.com. The site likely has a sitemap behind the Cloudflare wall, but it is inaccessible to crawlers.

    What to change: Ensure a sitemap.xml is served at a standard location and is accessible to AI crawlers.

11. Rich archived content is not served on live site High

    The Wayback Machine snapshot from March 2025 shows a full-featured site with inventory, service pages, financing, and a 7/30 Worry-Free Guarantee. None of this content is available on the live site to any automated agent.

    What to change: Restore the content from the archived version and serve it via server-side rendering so that AI crawlers can access it.

What's working

• Archived AutoDealer schema is well-structured — The March 2025 Wayback snapshot contains a complete AutoDealer schema with PostalAddress, OpeningHoursSpecification, and three departments (AutoPartsStore, AutoDealer, AutoRepair). This schema, if restored, would provide excellent structured data for AI models.
• Archived site had rich content including inventory and services — The archived homepage included inventory listings, service center details, financing options, and a 7/30 Worry-Free Guarantee. This content, if made accessible, would give AI models substantial information about the dealership.
• DealerRater reviews were embedded on archived homepage — The archived site displayed DealerRater reviews mentioning specific sales staff and service advisors. These reviews, if accessible, would provide social proof and third-party validation for AI models.
• DNS and SPF records are properly configured — The domain has valid DNS records (A, NS, TXT) and SPF records referencing CDK Global hosting and Google Workspace, indicating a professionally managed infrastructure.