tentbox.com — AI Site Grade

Analysis

Cloudflare Under Attack Mode Blocks All AI Crawlers

Every AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, and ChatGPT-User — receives a 403 response from tentbox.com. The site is behind Cloudflare's "Under Attack" mode, which serves a JS challenge page ("Verifying your connection...") to every request, including all bot user-agents tested. No bot receives any real content. The robots.txt and llms.txt endpoints also return 403 with the same JS challenge wall, making them inaccessible. The DNS records confirm the site is a Shopify store (via shopify-verification-code TXT record) hosted at 23.227.38.65 (Shopify/Cloudflare).

Cold-Knowledge Gap Is Extreme

The LLM model knows Tentbox as a UK-based roof-top tent brand with three product lines (Lite, Classic, Duo), a "pop-up" mechanism that sets up in under 60 seconds, and mixed reviews on durability and customer service. This knowledge likely comes from third-party reviews and social media, not from the site itself. The site's actual homepage, product pages, and about page are completely invisible to AI models. The model describes the brand's reputation as having "mixed reviews on durability and customer service" — a signal that external sources (Reddit, forums) dominate the brand's AI footprint more than its own content.

Zero Schema, Zero Structured Content

No JSON-LD schema was detected on any accessible page. The homepage returns only 10 words of visible text — the Cloudflare challenge message. No FAQ markup, no product schema, no breadcrumb, no Organization schema. The heading structure is a single <h1> reading "Your connection needs to be verified before you can proceed." There are no internal or external links extractable. The site has no llms.txt file accessible. The sitemap.xml also returns 403, meaning no URL discovery is possible for crawlers.

External Signals Dominate AI Perception

The DNS records reveal the brand has verified with Anthropic (anthropic-domain-verification-03t0sg), Apple (apple-domain-verification), and Google (google-site-verification), indicating the brand has taken steps to authenticate with these platforms. However, the Cloudflare Under Attack mode negates any benefit — verified crawlers still hit the JS wall. The brand's AI reputation is shaped entirely by off-domain sources: reviews, Reddit threads, and social media posts that the LLM model references in its cold knowledge. The gap between what the site could communicate (product specs, durability claims, warranty info) and what the model knows (mixed reviews, "pop-up" mechanism) represents a significant loss of narrative control.

Findings

1. Cloudflare Under Attack mode blocks all AI crawlers with 403 High

   Every AI crawler tested (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, ChatGPT-User) receives a 403 response with a JS challenge page. No bot receives real content.

   What to change: Disable Cloudflare Under Attack mode for known bot user-agents, or configure a firewall rule to allow AI crawlers through.

2. robots.txt returns 403, blocking crawler directives High

   The robots.txt endpoint returns a 403 JS challenge page, making it impossible for crawlers to read any crawl directives.

   What to change: Ensure robots.txt is publicly accessible without JS challenge, and include directives for AI crawlers.

3. Sitemap.xml returns 403, preventing URL discovery High

   The sitemap.xml endpoint returns 403, so crawlers cannot discover site URLs via sitemap.

   What to change: Make sitemap.xml publicly accessible and submit it to search engines.

4. No llms.txt file accessible Medium

   The llms.txt endpoint returns 403, so AI models have no structured guidance for site content.

   What to change: Create and serve an llms.txt file with a summary of the site's content and key pages.

5. No JSON-LD schema detected on any accessible page High

   No structured data (FAQ, Product, Organization, Breadcrumb) was found. The only accessible page is the Cloudflare challenge, which has no schema.

   What to change: Add JSON-LD schema markup for Organization, Product, and FAQ on relevant pages.

6. Homepage returns only Cloudflare challenge text High

   The homepage returns only 10 words of visible text ('Verifying your connection...'), with no product content, headings, or links.

   What to change: Disable Under Attack mode so the actual homepage content is served to crawlers.

7. Product and collection pages return 403 High

   URLs like /collections, /products, and /pages/about all return 403, making product catalog invisible to crawlers.

   What to change: Allow crawler access to product and collection pages by disabling Under Attack mode for those paths.

8. AI brand perception shaped entirely by external reviews Medium

   LLM cold knowledge describes Tentbox as having 'mixed reviews on durability and customer service' from third-party sources, while the site itself contributes no content to AI models.

   What to change: Make site content accessible to AI crawlers to regain narrative control over product quality and warranty information.

9. No internal or external links extractable from homepage Medium

   The Cloudflare challenge page contains no links, so crawlers cannot discover other pages through link traversal.

   What to change: Serve actual homepage content with navigation links to allow crawler discovery.

10. Zero URLs discovered via known URL list Medium

    The list_known_urls tool returned 0 URLs, indicating no public index of site pages.

    What to change: Ensure sitemap.xml is accessible and submit it to search engines.

11. Web searches for tentbox.com return zero results High

    Multiple web searches for the domain and brand name returned no results, indicating poor search engine indexing.

    What to change: Resolve Cloudflare blocking to allow search engine crawlers to index the site.

12. Wayback Machine snapshots return connection errors Low

    Archived snapshots of the homepage and robots.txt return connection refused errors, suggesting the site has been blocking for some time.

    What to change: Investigate server configuration to allow Wayback Machine access.

What's working

• Domain verified with Anthropic, Apple, and Google — DNS records show verification tokens for Anthropic, Apple, and Google, indicating the brand has taken steps to authenticate with these platforms.
• Shopify hosting identified via DNS — DNS records confirm the site is hosted on Shopify (IP 23.227.38.65) with a Shopify verification TXT record, providing a clear hosting context.