maeving.com — AI Site Grade

Analysis

The key findings: 1. The entire live site is behind Cloudflare JS challenge — every AI bot AND every browser gets 403 2. The site runs on Shopify (evident from robots.txt and DNS) 3. No llms.txt exists (gets Cloudflare-blocked) 4. robots.txt exists but has NO AI-specific bot rules — no GPTBot, ClaudeBot, PerplexityBot, Google-Extended directives at all 5. The cold LLM knowledge knows Maeving well (RM1 details, Coventry, retro styling) but is missing the RM1S and RM2 models entirely 6. The help centre (help.maeving.com) on HubSpot IS accessible to crawlers and contains rich FAQ content 7. DuckDuckGo returns ZERO search results for the brand — total invisibility 8. The site has a redirect from /pages/maeving-rm1-electric-motorcycle to /en-gb/pages/maeving-rm1-electric-motorcycle (locale prefix) 9. Schema is minimal — only Organization and WebSite types, no Product schema on product pages 10. The RM2 model exists (mentioned in help centre) but has no product page on the main site visible in the Wayback snapshot

The single most non-obvious finding: The site's Cloudflare challenge wall blocks ALL traffic — including real browsers — meaning even Google's crawler cannot index the live site. The only reason the site appears in any index at all is historical Wayback captures.

Let me write the audit.

The Cloudflare JS challenge wall on maeving.com blocks every AI crawler and every browser alike — the live site is functionally invisible to all search and AI engines.

Crawler Access

Every AI crawler tested — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Applebot-Extended, Bytespider, anthropic-ai, ChatGPT-User, Perplexity-User — receives a 403 Cloudflare JS challenge wall identical to the browser baseline. The robots.txt at /robots.txt is also behind the same wall (though a Wayback capture shows it exists with standard Shopify boilerplate). Critically, the robots.txt contains no AI-bot-specific directives — no User-agent: GPTBot, no User-agent: ClaudeBot, no User-agent: Google-Extended. The llms.txt returns the same 403 wall. The site runs on Shopify (A record 23.227.38.65, Shopify boilerplate in robots.txt) behind Cloudflare with Under Attack mode or equivalent JS challenge enabled.

Cold-Knowledge Gap

The cold LLM knows Maeving as a Coventry-based British manufacturer of the RM1, a retro-styled 45 mph city bike with removable batteries. This is accurate but stale by at least one product generation. The live site (via Wayback) prominently sells the RM1S (70 mph, 80-mile range, from £7,495) and the help centre references an RM2 model — neither exists in the model's prior knowledge. The RM1S is the flagship, yet the LLM has no awareness of it. The brand's positioning as "award-winning Electric Bike of the Year" (MCN award, back-to-back) is also absent from cold knowledge.

Schema Posture

The homepage carries only Organization and WebSite schema types. The RM1S product page (via Wayback) also carries only Organization schema — no Product schema, no Offer, no Vehicle schema, no FAQPage despite having a FAQ section with 7 questions. No breadcrumb, no review aggregate schema despite 8 press quotes displayed. The help centre (help.maeving.com, hosted on HubSpot) carries zero structured data across any of its 16+ knowledge base sections.

External Signals

DuckDuckGo returns zero search results for "Maeving", "Maeving RM1", "Maeving RM1S", or "Maeving electric motorcycle" — total organic invisibility on that engine. The brand's external web footprint is effectively limited to its own domain and the HubSpot help centre. The Wayback Machine has 138 captures of robots.txt since November 2021, indicating the site has existed for years but the Cloudflare wall appears to be a recent or persistent hardening. The help centre at help.maeving.com is the only publicly crawlable content — it contains rich FAQ data about all three models (RM1, RM1S, RM2) but is entirely disconnected from the main site's schema and has no internal linking from the product pages.

Findings

1. Cloudflare JS challenge wall blocks every AI crawler and browser High

   All 11 tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, etc.) receive a 403 Cloudflare JS challenge, identical to browser traffic. The live site is functionally invisible to all search and AI engines.

   What to change: Disable Cloudflare Under Attack mode or JS challenge for known bot user agents. Allow GPTBot, ClaudeBot, Google-Extended, and other AI crawlers through via Cloudflare WAF rules or robots.txt allow directives.

2. Robots.txt contains no AI-bot-specific directives High

   The robots.txt file (visible via Wayback) has standard Shopify boilerplate but no User-agent lines for GPTBot, ClaudeBot, Google-Extended, or any other AI crawler. Combined with the Cloudflare wall, AI bots have no path to access the site.

   What to change: Add explicit User-agent directives for GPTBot, ClaudeBot, Google-Extended, and other AI crawlers to allow crawling of key pages.

3. No llms.txt file accessible Medium

   The llms.txt endpoint returns a 403 Cloudflare wall, making it impossible for AI agents to discover structured site information.

   What to change: Create an llms.txt file with a summary of the site, key pages, and links to product documentation. Ensure it is not blocked by Cloudflare.

4. Cold LLM knowledge missing RM1S and RM2 models High

   The cold LLM knows only the RM1 model (45 mph city bike). The RM1S (70 mph, £7,495) and RM2 are absent from prior knowledge, despite being the current flagship and a newer model respectively.

   What to change: Ensure product pages for RM1S and RM2 are crawlable and contain rich structured data (Product schema, descriptions, specs) so LLMs can index them.

5. No Product schema on product pages High

   The RM1S product page (via Wayback) carries only Organization schema. No Product, Offer, Vehicle, or FAQPage schema exists despite a FAQ section and pricing information.

   What to change: Add Product schema with name, description, brand, offers (price, currency, availability), and Vehicle schema for motorcycle attributes. Add FAQPage schema for the FAQ section.

6. FAQ section on product page lacks FAQPage schema Medium

   The RM1S page has a FAQ section with 7 questions but no FAQPage structured data, missing an opportunity for rich results and AI extraction.

   What to change: Implement FAQPage schema for the FAQ section on product pages.

7. Help centre pages have zero structured data Medium

   The HubSpot-hosted help centre (help.maeving.com) contains rich FAQ content about all three models but no structured data of any kind, limiting AI extraction.

   What to change: Add FAQPage, Article, or QAPage schema to help centre articles, especially those covering product details.

8. Zero DuckDuckGo search results for the brand Medium

   DuckDuckGo returns no results for queries like 'Maeving', 'Maeving RM1', 'Maeving RM1S', or 'Maeving electric motorcycle', indicating total organic invisibility on that engine.

   What to change: Improve overall crawlability and indexing by removing the Cloudflare wall, then build external backlinks and citations to increase search engine visibility.

9. RM2 model has no product page on main site Medium

   The RM2 model is mentioned in the help centre but has no corresponding product page on the main site (visible in Wayback), creating a gap in product information for AI crawlers.

   What to change: Create a dedicated product page for the RM2 on the main site with full specifications, pricing, and structured data.

10. No breadcrumb schema on product pages Low

    The RM1S product page lacks breadcrumb structured data, which helps search engines understand site hierarchy.

    What to change: Add BreadcrumbList schema to product pages.

11. No Review aggregate schema despite press quotes Low

    The RM1S page displays 8 press quotes but no AggregateRating or Review schema, missing an opportunity to showcase awards and reviews in rich results.

    What to change: Add AggregateRating schema with rating value and review count based on press quotes and awards.

12. Help centre not linked from main site product pages Medium

    The help centre (help.maeving.com) is the only crawlable content but has no internal links from the main site's product pages, creating a silo that reduces its discoverability.

    What to change: Add links from product pages to relevant help centre articles, and ensure the help centre is crawlable and indexed.

What's working

• Cold LLM knowledge accurately describes RM1 model — The cold LLM correctly identifies Maeving as a Coventry-based British manufacturer of the RM1, a retro-styled 45 mph city bike with removable batteries, indicating some baseline brand awareness.
• Help centre is crawlable and contains rich FAQ content — The HubSpot help centre at help.maeving.com is accessible to crawlers and contains detailed FAQ content about all three models (RM1, RM1S, RM2), providing a source of structured information.
• Wayback Machine captures preserve site content — The Wayback Machine has snapshots of the main site and robots.txt, providing historical content that can be used for recovery and analysis.
• Organization schema present on homepage — The homepage includes Organization and WebSite schema types, providing basic brand information to search engines.
• Shopify platform with standard setup — The site runs on Shopify, a well-supported e-commerce platform with built-in SEO features and structured data capabilities.
• Locale redirect for product pages — The site redirects /pages/maeving-rm1-electric-motorcycle to /en-gb/pages/maeving-rm1-electric-motorcycle, indicating locale handling for international audiences.
• Award mentions on product page — The RM1S product page displays 'Electric Bike of the Year' awards from MCN, providing social proof that could be leveraged with structured data.
• DNS and hosting configuration stable — The domain has consistent DNS records (A record to Shopify, NS, TXT) and Cloudflare protection, indicating a professionally managed infrastructure.