martingp.com — AI Site Grade

Analysis

Anthropic-ai Gets Full Content While Every Other Crawler and Browser Gets 403

martingp.com, the digital storefront for Martin Management Group — a multi-franchise auto dealership group in Bowling Green, KY — is effectively invisible to the open web. The site serves a 403 "Access Denied" page to every browser and every major AI crawler (GPTBot, Google-Extended, ClaudeBot, PerplexityBot, ChatGPT-User, OAI-SearchBot, Applebot-Extended, Bytespider) via an Akamai (edgesuite.net) WAF. The sole exception: anthropic-ai receives a full 200 response with 868KB of rendered HTML, served from an nginx backend.

Crawler Access

The robots.txt explicitly blocks GPTBot, OAI-SearchBot, ChatGPT-User, Claude-User, Claude-SearchBot, and PerplexityBot from /api/, /apis/, /pixall/, and all .js, .css, and .json resources — but does not block them from the root. This is irrelevant, because the Akamai WAF blocks those bots at the edge before they ever reach the robots.txt rules. The anthropic-ai bot bypasses the WAF entirely, landing on a different server (nginx vs AkamaiGHost). No llms.txt exists (connection refused). The sitemap.xml (74KB, containing hundreds of inventory and content URLs) is also only accessible to anthropic-ai.

Cold-Knowledge Gap

LLM knowledge of "Martin GP" or "Martin Management Group" is zero. The model reports no verifiable information about the brand, its products, or its reputation. This is a complete vacuum — the site's content, which includes a 40-year history, 13+ automotive franchises, and a "Rev Rewards" loyalty program, is entirely absent from AI training data. The gap between what the site contains and what AI models know is total.

Content and Schema

The homepage title lists 13 car brands (Volkswagen, Kia, Dodge, Jeep, Subaru, Audi, Mitsubishi, Chrysler, Honda, Ford, Hyundai, Ram, Nissan) and positions the group as a Bowling Green, KY dealership. The about page and 40th-anniversary page exist but their full visible text could not be extracted due to the JS-heavy DDC (Dealer.com) platform architecture. The site carries meta name="robots" content="index, follow" on interior pages, but the WAF blocks Googlebot from ever rendering them. No JSON-LD schema was detected in the truncated HTML snippets — the site relies on the DDC platform's default markup, which typically includes AutoDealer schema but may not be served to blocked crawlers.

External Signals

No external mentions, reviews, Reddit threads, or press coverage were found for "Martin Management Group" or martingp.com across web search. The business operates entirely outside the indexed web. The domain's DNS (GoDaddy nameservers, Mimecast email) and IP (64.70.56.99) show no other hosted content. The site's lastmod dates in the sitemap show 2026-05-30, suggesting either a future-dated cache or a misconfigured timestamp.

Findings

1. Akamai WAF blocks all browsers and most AI crawlers with 403 High

   The site returns a 403 Access Denied page to every standard browser and every major AI crawler (GPTBot, Google-Extended, ClaudeBot, PerplexityBot, ChatGPT-User, OAI-SearchBot, Applebot-Extended, Bytespider) via an Akamai WAF. Only anthropic-ai receives a full 200 response.

   What to change: Reconfigure the Akamai WAF to allow standard browsers and all major AI crawlers (Googlebot, GPTBot, ClaudeBot, etc.) to access the site, or remove the WAF entirely.

2. Anthropic-ai bot bypasses WAF and reaches different backend High

   The anthropic-ai bot receives a 200 response from an nginx backend, while all other traffic hits an AkamaiGHost server that returns 403. This selective access creates an uneven AI visibility landscape.

   What to change: Ensure consistent access policies across all crawlers and backends, or document the intentional bypass.

3. Robots.txt blocks GPTBot, OAI-SearchBot, and others from resources Medium

   The robots.txt explicitly disallows GPTBot, OAI-SearchBot, ChatGPT-User, Claude-User, Claude-SearchBot, and PerplexityBot from accessing /api/, /apis/, /pixall/, and all .js, .css, and .json files. While the WAF makes this moot for most crawlers, it further restricts access for any that might bypass the WAF.

   What to change: Remove disallow rules for AI crawlers from robots.txt, or at minimum allow access to necessary resources.

4. No llms.txt file available Medium

   The llms.txt endpoint at martingp.com returns a connection refused error, meaning no guidance is provided to AI crawlers about which content to use.

   What to change: Create an llms.txt file that lists key pages and content for AI crawlers.

5. Sitemap.xml only accessible to anthropic-ai High

   The sitemap.xml (74KB, containing hundreds of URLs) returns 403 when fetched by a standard HTTP client, but is accessible to anthropic-ai. This prevents search engines and other crawlers from discovering the site's content.

   What to change: Make sitemap.xml publicly accessible to all crawlers.

6. LLMs have zero knowledge of Martin Management Group High

   Web search and LLM knowledge retrieval return no results for 'Martin Management Group' or 'martingp.com'. The brand, its 13 franchises, 40-year history, and loyalty program are entirely absent from AI training data.

   What to change: Allow crawlers to index the site and consider submitting business information to knowledge panels and directories.

7. No external web presence or backlinks found Medium

   Web searches for the dealership group, its domain, and related terms return zero results. There are no reviews, press articles, or directory listings indexed.

   What to change: Build external citations through Google Business Profile, Yelp, DealerRater, and local press.

8. No JSON-LD schema detected in truncated HTML Medium

   The HTML snippets retrieved from interior pages do not contain JSON-LD structured data. The site likely relies on DDC platform defaults, but schema may not be served to blocked crawlers.

   What to change: Implement JSON-LD schema (AutoDealer, LocalBusiness, Product) on all pages and ensure it is served to all crawlers.

9. JS-heavy DDC platform limits content extraction Medium

   The site uses the Dealer.com (DDC) platform, which renders content via JavaScript. Full visible text could not be extracted from interior pages, potentially hiding content from crawlers that do not execute JS.

   What to change: Ensure critical content is server-side rendered or available in static HTML for crawlers that do not execute JavaScript.

10. Sitemap contains future-dated lastmod timestamps Low

    The sitemap shows lastmod dates of 2026-05-30, which is in the future. This may confuse crawlers or indicate a misconfigured cache.

    What to change: Correct the lastmod dates in the sitemap to reflect actual last modification times.

11. No Wayback Machine snapshots exist Low

    The domain has no archived snapshots in the Wayback Machine, indicating it has never been publicly accessible or indexed historically.

    What to change: Ensure the site is publicly accessible to allow archiving and historical indexing.

What's working

• Anthropic-ai bot receives full 200 response with rich content — The anthropic-ai bot successfully retrieves 868KB of rendered HTML from the homepage, including inventory listings and brand information. This ensures at least one AI model can access the site's content.
• Sitemap.xml accessible to anthropic-ai with hundreds of URLs — The sitemap.xml (74KB) is served to anthropic-ai, containing hundreds of inventory and content URLs that could be crawled.
• Robots.txt does not block anthropic-ai — The robots.txt file does not contain any disallow rules for anthropic-ai, allowing it to crawl the site freely.
• Interior pages include meta robots index, follow — The about page and other interior pages contain <meta name="robots" content="index, follow">, indicating an intent to be indexed.
• Homepage clearly lists 13 automotive franchises — The homepage title enumerates 13 car brands (Volkswagen, Kia, Dodge, Jeep, Subaru, Audi, Mitsubishi, Chrysler, Honda, Ford, Hyundai, Ram, Nissan), providing clear brand identity.
• DNS and email infrastructure are properly configured — The domain has valid A, NS, and TXT records, including Mimecast email security, indicating operational infrastructure.