courtesycars.com — AI Site Grade

Analysis

Anthropic-Only Access: The Akamai Firewall That Lets One Bot Through

The site is invisible to every major AI crawler except Anthropic's own. GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User, and Applebot-Extended all receive HTTP 403 from AkamaiGHost on the homepage, while anthropic-ai gets a 200 with 278KB of real HTML served from nginx. This is not a robots.txt issue — the file itself returns 403 — but an edge-firewall rule that explicitly whitelists one bot user-agent and blocks everything else, including standard browsers.

Crawler Access

Every tested bot except anthropic-ai is blocked at the Akamai edge with a generic "Access Denied" page. The robots.txt and llms.txt files are also 403'd, meaning no crawler can even read access directives. The sitemap.xml returns 403. Google Search returns zero indexed pages for site:courtesycars.com. The site runs on Dealer.com (DDC) platform behind Cloudflare DNS and Akamai CDN, with nginx serving the anthropic-ai traffic. The anthropic-ai user-agent is the only known path to the site's content.

Cold-Knowledge Gap

LLM knowledge describes Courtesy Cars as a standalone "buy here, pay here" used-car lot specializing in subprime financing with vehicles under $15,000. The actual site reveals a nine-dealership, twelve-brand auto group (Chrysler, Dodge, Hyundai, Mercedes-Benz, Honda, Infiniti, Jeep, Kia, Nissan, Ram, Toyota, Volvo) with new and used inventory, factory-certified vehicles, online sales via Clicklane, and a stated inventory of 10,000+ vehicles. The cold knowledge describes a small subprime operation; the site describes a major multi-franchise group owned by a larger entity (Asbury Automotive Group, based on the lhm picture subdomain and landcar.com cancellation link). The gap is a complete category mismatch.

Schema Posture

The homepage and key pages contain no JSON-LD structured data visible in the HTML head. The Wayback Machine snapshot confirms the live site has no AutoDealer, Organization, Product, or Review schema markup. The meta name="robots" content="index, follow" tags are present on interior pages, but they are irrelevant when Akamai blocks all search engine crawlers at the edge. The site has google-site-verification and msvalidate.01 meta tags, indicating past SEO setup, but the firewall renders them inert.

External Signals

The brand has zero indexed pages on Google and zero search results for its domain name or dealership group name in web searches. The Wayback Machine shows 701 captures since 1999, confirming a long-established site that has recently become firewall-locked. The customer-reveiws.htm page (note the typo in the URL) exists but is also only accessible to anthropic-ai. No external review sites, Reddit threads, or press mentions surfaced in searches — the domain is effectively air-gapped from the open web.

Findings

1. Akamai firewall blocks all AI crawlers except Anthropic High

   Every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User, Applebot-Extended) receives HTTP 403 from AkamaiGHost on the homepage, while anthropic-ai gets a 200 with full HTML. Standard browsers also get 403.

   What to change: Remove the Akamai edge firewall rule that blocks non-anthropic-ai user-agents, or replace it with a rate-limited allowlist for all major AI crawlers and search engine bots.

2. Robots.txt and sitemap.xml return 403 High

   The robots.txt file, llms.txt, and sitemap.xml all return HTTP 403, preventing any crawler from reading access directives or discovering URLs.

   What to change: Serve robots.txt and sitemap.xml publicly with HTTP 200, and include directives that allow all major AI crawlers.

3. Zero pages indexed on Google High

   Google Search returns zero results for site:courtesycars.com, and no URLs were discovered via known URL lists. The site is completely absent from search engine indexes.

   What to change: Allow Googlebot through the firewall, submit a sitemap to Google Search Console, and ensure all pages are crawlable and indexable.

4. No JSON-LD structured data on any page High

   The homepage and interior pages contain no JSON-LD schema markup for AutoDealer, Organization, Product, or Review. The Wayback Machine snapshot confirms the absence of structured data.

   What to change: Add JSON-LD structured data for AutoDealer, Organization, Vehicle, and Review to all relevant pages.

5. Cold knowledge misrepresents dealership as small subprime lot High

   LLM knowledge describes Courtesy Cars as a standalone 'buy here, pay here' used-car lot, but the site reveals a nine-dealership, twelve-brand auto group with 10,000+ vehicles. This mismatch harms AI-generated summaries.

   What to change: Publish a comprehensive 'about' page with accurate business description, and ensure structured data reflects the full dealership group scope.

6. No external web presence or reviews found Medium

   Web searches for the dealership name, reviews, and group affiliation returned zero results. The domain is effectively air-gapped from the open web.

   What to change: Build citations on third-party review sites, local directories, and social media to establish an external web presence.

7. Standard browsers blocked by Akamai High

   Regular fetch requests to the homepage and interior pages return HTTP 403, meaning human visitors are also blocked unless they use the anthropic-ai user-agent.

   What to change: Allow standard browser user-agents through the Akamai firewall, or implement a proper bot detection system that does not block all browsers.

8. Typo in reviews page URL Low

   The reviews page URL contains a typo: 'customer-reveiws.htm' instead of 'reviews'. This may cause broken links and confusion.

   What to change: Fix the typo in the URL to 'customer-reviews.htm' and set up a redirect from the old URL.

What's working

• Anthropic AI crawler receives full HTML content — The anthropic-ai user-agent successfully retrieves 278KB of real HTML from the homepage and interior pages, indicating the site's content is accessible to at least one AI crawler.
• Google and Bing verification meta tags present — The homepage includes google-site-verification and msvalidate.01 meta tags, indicating past SEO setup and ownership verification.
• Robots meta tags allow indexing on interior pages — Interior pages contain <meta name="robots" content="index, follow">, which would be effective if the firewall allowed crawlers.
• Long history of Wayback Machine captures — The domain has 701 captures since 1999, indicating a long-established site with historical content.