keatingauto.com — AI Site Grade

Analysis

Cloudflare Blocks All Browsers and Most AI Crawlers — Only OpenAI and Perplexity User-Agents Get Through

The site at keatingauto.com is a used-car dealership running on the Dealer eProcess platform (dealereprocess.org), but Cloudflare's WAF is configured to block every browser request and most AI crawlers while selectively allowing three bot user-agents through. This creates a bizarre visibility funnel where only ChatGPT-User, OAI-SearchBot, and Perplexity-User receive a 200 with full HTML content — every other visitor, including GPTBot, ClaudeBot, Google-Extended, PerplexityBot, and all real human browsers, gets a 403 Cloudflare block page.

Crawler Access

The compare_bot_access test on the homepage returned 403 for 9 of 12 user-agents tested. The only passes were ChatGPT-User (200, 445KB), OAI-SearchBot (200, 445KB), and Perplexity-User (200, 447KB). GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Bytespider, anthropic-ai, and a standard browser all hit the same Cloudflare block. The robots.txt and llms.txt endpoints both return the Cloudflare 403 HTML page — neither file exists or is served. The sitemap at /sitemap.xml also 403s. The site uses Cloudflare DNS (jeff.ns.cloudflare.com, tricia.ns.cloudflare.com) with A records pointing to Cloudflare IPs (104.17.x.x).

Content and Schema

The homepage title reads "Keating Auto Group - Used & Pre-Owned Vehicles near Port Lavaca, TX" with a meta description targeting College Station and Victoria, TX. The page is 445KB of HTML — almost entirely JavaScript bundles from cdn.dealereprocess.org — making it a heavy JS-rendering shell. The homepage body snippet shows no visible text content, no headings, no JSON-LD schema, and no structured data of any kind. The /inventory, /about, and /contact paths all return 404 "Page Not Found" pages. The only working page is the homepage, and it delivers zero crawlable text.

Cold-Knowledge Gap

The LLM model (deepseek-v4-flash) describes Keating Auto Group as a Houston-area network of dealerships selling Ford, Lincoln, Toyota, and Nissan — a family-owned group with decades of history. The actual site says nothing about Houston, makes no mention of Ford/Lincoln/Toyota/Nissan, and describes itself as serving Port Lavaca, College Station, and Victoria — three small Texas cities 100-200 miles from Houston. The cold knowledge describes a multi-brand new-car dealer group; the site presents as a single-location used/pre-owned lot. This is a fundamental identity mismatch: the AI models know a different business than the one the domain actually represents.

External Signals

Zero external signals were found. Web searches for "Keating Auto Group," "Keating Auto Texas," and related queries returned no results across Google, news, reviews, social media, or dealer directories. No Wayback Machine snapshots exist. No Reddit threads, press mentions, or review pages surfaced. The domain has no discoverable off-domain footprint whatsoever — a striking absence for any operating dealership.

Findings

1. Cloudflare blocks all browsers and most AI crawlers High

   The site returns 403 for 9 of 12 tested user-agents, including GPTBot, ClaudeBot, Google-Extended, and standard browsers. Only ChatGPT-User, OAI-SearchBot, and Perplexity-User receive 200 responses.

   What to change: Configure Cloudflare WAF to allow all legitimate AI crawlers and human browsers, or use a more permissive security approach.

2. No robots.txt, llms.txt, or sitemap accessible High

   The robots.txt, llms.txt, and sitemap.xml endpoints all return 403 Cloudflare block pages. No crawl directives or URL lists are available.

   What to change: Create and serve a robots.txt that allows AI crawlers, an llms.txt for LLM guidance, and a valid sitemap.xml.

3. Homepage is a heavy JS shell with zero crawlable text High

   The 445KB homepage contains almost entirely JavaScript bundles from cdn.dealereprocess.org. No visible text, headings, or structured data is present in the raw HTML.

   What to change: Implement server-side rendering or pre-rendering to deliver meaningful HTML content to crawlers.

4. Key pages return 404 for AI crawlers High

   The /inventory, /about, and /contact paths all return 404 pages when fetched as ChatGPT-User. These pages are inaccessible to crawlers.

   What to change: Ensure these pages exist and return 200 with crawlable content, or redirect to the homepage.

5. No JSON-LD or structured data on the homepage High

   The homepage HTML contains no JSON-LD schema, no structured data for LocalBusiness, AutoDealer, or any other type. AI crawlers cannot extract entity information.

   What to change: Add JSON-LD structured data for AutoDealer, LocalBusiness, and VehicleInventory, including name, address, phone, and inventory details.

6. AI cold knowledge describes a different business than the site Medium

   LLM knowledge describes Keating Auto Group as a Houston-area multi-brand new-car dealer (Ford, Lincoln, Toyota, Nissan). The site presents as a single-location used-car lot in Port Lavaca, TX. This mismatch undermines AI trust.

   What to change: Align site content with the actual business identity, or update external listings to reflect the correct locations and brands.

7. Zero external signals found across web searches High

   Web searches for the dealership name, domain, and related queries returned no results. No Google index, reviews, social media, or directory listings were found. The domain has no off-domain footprint.

   What to change: Build external signals by claiming Google Business Profile, listing on dealer directories, and encouraging reviews on third-party sites.

8. No Wayback Machine snapshots exist Low

   The Wayback Machine has no snapshots for keatingauto.com, indicating the domain may be new or has never been archived.

What's working

• OpenAI and Perplexity user-agents are allowed and served full HTML — ChatGPT-User, OAI-SearchBot, and Perplexity-User receive 200 responses with full HTML content, enabling some AI visibility.
• Homepage has a descriptive title and meta description — The title and meta description mention the business name and target cities (Port Lavaca, College Station, Victoria), providing basic context.