tetontoyota.com — AI Site Grade

Analysis

I have enough data to write a thorough audit. Let me compile.

The llms.txt endpoint silently redirects to the homepage (200, no error), meaning the site has no AI-friendly content map at all — and the cold-knowledge model already knows about Teton Auto Credit and Teton Automotive Group, neither of which the homepage prominently features.

Crawler Access

The robots.txt at tetontoyota.com is unusually comprehensive for a local dealership. It explicitly names GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, Perplexity-User, Google-Extended, Applebot, Bytespider, and Amazonbot — all granted Allow: / with only narrow disallows on /inventory*,*, /login, /print, /scheduletestdrive?*, and /reserve-your-vehicle/. Every AI bot tested returns 200 with full content (354 KB, matching browser baseline). The site runs on Fastly CDN with Signal Sciences (sigsci) WAF and Kestrel server — no Cloudflare, no JS-challenge, no UA-based blocking. Inventory listing pages like /inventory/new/toyota/tacoma also return 785 KB of HTML to GPTBot and ClaudeBot with no truncation.

Cold-Knowledge Gap

The LLM's prior knowledge describes Teton Toyota as "part of the Teton Automotive Group, which operates multiple dealerships in the region" and mentions a "showroom near the Snake River." The actual site never mentions Teton Automotive Group on any fetched page. The homepage says "Since 2005" but the model's prior has no founding-year awareness. The model also recalls "mixed reviews with praise for sales staff but occasional complaints about service wait times" — the site itself contains zero customer reviews, no testimonials, no Google/DealerRater widget, and no reputation management content. The Teton Auto Credit sub-brand (buy-here-pay-here financing) is a major differentiator on the site but the cold model only vaguely references financing.

Schema Posture

Every page carries a complete AutoDealer JSON-LD block with PostalAddress, GeoCoordinates, openingHoursSpecification, sameAs (Facebook and Twitter), and contactPoint. The inventory pages additionally embed per-vehicle Vehicle schema with VIN, price, Offers, and itemCondition. However, the sameAs array is stale: it links to twitter.com/tetontoyota?lang=en (X/Twitter, not the current platform name) and omits Instagram, despite the homepage text explicitly mentioning an Instagram page. The priceRange value $1200-$15000 appears low for a full-line Toyota dealer selling Tundras and Sequoias above $80K.

External Signals

Web searches for "Teton Toyota Idaho Falls reviews" and Reddit mentions returned zero results — the site has essentially no discoverable third-party citation footprint. The DNS TXT records include SFMC (Salesforce Marketing Cloud) and apple-domain-verification, indicating a modern martech stack. The Wayback Machine has no snapshot of the domain, suggesting the site may have been rebuilt or the domain changed hands. The sitemap index contains 6 sub-sitemaps with hundreds of inventory VDP URLs, but the pagelinkssitemap.xml and researchmodelsitemap.xml suggest thin content pages that may not carry unique value for AI crawlers.

Surprising Details

The /llms.txt URL returns a 200 with the homepage HTML (no redirect, no 404) — the server silently serves the homepage for any unknown path, which means AI crawlers hitting /llms.txt get a full dealership homepage instead of a structured content map. The inventory JSON feeds (/inventory-new.json, /inventory-used.json, /inventory-cpo.json) are publicly accessible and contain rich structured data (VIN, MSRP, selling price, MPG, drivetrain) that no AI crawler is directed to. The homepage has no FAQ schema, no comparison tables, and no answer-format signals despite being a natural candidate for "Toyota Tacoma vs Tundra" or "new vs used" comparison content. The finance page mentions "Coos Bay" (an Oregon city 600+ miles away) in a copy-paste error from a template.

Findings

1. llms.txt endpoint silently serves homepage HTML High

   The /llms.txt URL returns a 200 with the full homepage HTML instead of a structured content map, meaning AI crawlers get no AI-friendly content summary.

   What to change: Serve a proper llms.txt file with links to key pages and data feeds.

2. Zero third-party review citations found High

   Web searches for 'Teton Toyota Idaho Falls reviews' and Reddit mentions returned zero results, indicating no discoverable third-party citation footprint.

   What to change: Encourage customers to leave reviews on Google, DealerRater, and Reddit; embed a review widget on the site.

3. Stale sameAs links in JSON-LD Medium

   The AutoDealer schema's sameAs array links to Twitter (X) but omits Instagram, despite the homepage mentioning an Instagram page.

   What to change: Update sameAs to include Instagram and rename Twitter to X if appropriate.

4. priceRange value appears too low Medium

   The priceRange in AutoDealer schema is $1200-$15000, which is low for a full-line Toyota dealer selling vehicles above $80K.

   What to change: Update priceRange to reflect the actual range of vehicles sold, e.g., $1200-$85000.

5. Finance page contains copy-paste error referencing Coos Bay Medium

   The finance department page mentions 'Coos Bay', an Oregon city 600+ miles away, indicating a template error.

   What to change: Remove the erroneous 'Coos Bay' reference from the finance page.

6. No FAQ schema or comparison content on homepage Medium

   The homepage lacks FAQ schema and comparison tables (e.g., Tacoma vs Tundra), missing opportunities for AI answer extraction.

   What to change: Add FAQ schema for common questions and create comparison pages for popular models.

7. Cold LLM knowledge does not match site content Medium

   The LLM's prior knowledge mentions Teton Automotive Group and mixed reviews, but the site never mentions the group and has no reviews.

   What to change: Add a section about Teton Automotive Group and embed customer reviews on the site.

8. No Wayback Machine snapshot available Low

   The Wayback Machine has no snapshot of the domain, suggesting a recent rebuild or domain change.

   What to change: Ensure the site is archived by submitting to the Wayback Machine.

What's working

• Comprehensive robots.txt with explicit AI bot allowances — The robots.txt explicitly allows all major AI bots (GPTBot, ClaudeBot, etc.) with only narrow disallows, ensuring full crawler access.
• AI bots receive full HTML content — All tested AI bots (GPTBot, ClaudeBot) receive 785 KB of HTML for inventory pages, matching browser baseline with no truncation.
• Complete AutoDealer JSON-LD on every page — Every page carries a full AutoDealer schema with address, coordinates, hours, sameAs, and contactPoint.
• Per-vehicle Vehicle schema on inventory pages — Inventory pages embed Vehicle schema with VIN, price, Offers, and itemCondition for each vehicle.
• Publicly accessible inventory JSON feeds — Inventory JSON feeds (/inventory-new.json, etc.) are publicly accessible and contain rich structured data (VIN, MSRP, MPG).
• Modern martech stack with Salesforce Marketing Cloud — DNS TXT records include SFMC and apple-domain-verification, indicating a modern marketing technology stack.
• Fastly CDN with WAF protection — The site uses Fastly CDN and Signal Sciences WAF, providing performance and security without blocking AI crawlers.