promixx.com — AI Site Grade

Analysis

PROMIXX: AI crawlers see the storefront but not the catalog

The site's llms.txt and agents.md are among the most sophisticated agent-commerce documents on any Shopify store — they define a full Universal Commerce Protocol (UCP) with MCP endpoints, versioned API support, and explicit agent shopping flows — yet the actual product catalog is partially invisible to every AI crawler tested.

Crawler Access

All 10 AI bot UAs (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User) receive 200 with full content on the homepage and key product pages (Sense, Pro). The robots.txt is the standard Shopify template with no AI-bot-specific disallows — User-agent: * allows / broadly. However, many product pages — including /products/pursuit, /products/miixr-stealth, and /products/miixr-x7-electric-protein-shake-blender-usa — return 403 from Cloudflare to every UA including browser baseline. These are not blocked by robots.txt but by Cloudflare's JS challenge/WAF, making them invisible to all crawlers. The sitemap lists these URLs, so crawlers discover them, then hit a wall.

Cold-Knowledge Gap

A frontier LLM queried cold about PROMIXX describes the brand primarily through its iMix and iMix Pro products — battery-powered shakers with built-in impellers. The current site contains zero mention of "iMix" anywhere. The brand has fully rebranded its product line to Pro Series (electric shaker bottle, £30.99), Sense (cordless portable blender, £90.99), and Pursuit (manual shakers). The old Miixr+ Stealth and Miixr X7 product pages return 403. The LLM also references "self-cleaning" and "quiet operation" as differentiators — the current site emphasizes "X-Blade technology" and "Vortex Mixing Technology" instead. The cold knowledge is stuck on a discontinued product generation.

Schema Posture

Product pages carry rich JSON-LD: Product with ProductGroup variants, AggregateRating (Sense: 4.96/23 reviews, Pro: 4.92/600 reviews), gtin13, mpn, sku, priceValidUntil through 2026-08-28, and BreadcrumbList. The homepage has Organization and WebSite with SearchAction. Missing: FAQPage schema on the reviews page (which has actual FAQ content about usage and cleaning), and Review schema objects for individual testimonials. The blog uses Blog/ItemList schema but article pages lack Article or NewsArticle markup.

External Signals

The site is hosted on Shopify (23.227.38.65, Cloudflare edge) with Microsoft 365 mail. The reviews page links to press mentions in British GQ, T3, Men's Health UK, and Daily Mail — the GQ article (2023) references the now-discontinued "Miixr+ Stealth" at £49.99. The site claims "12,000+ 5-star reviews" and "5 Million Customers & 15,000 Stores Worldwide." The blog "ACADEMY" is actively publishing (latest post April 2026) with content about HYROX training, gut health, and high-protein snacking — a strong content marketing engine for AI retrieval.

Surprising Findings

The llms.txt and agents.md documents are ahead of the industry — they define UCP versions 2026-04-08 and 2026-01-23, MCP tool endpoints, and explicit agent shopping flows with buyer-approval requirements. This is Shopify's standard agent-commerce template, but its presence means PROMIXX is technically ready for AI-agent-driven purchasing. The irony: the agent infrastructure is excellent, but the product catalog has a Cloudflare 403 problem on many SKUs. The old product URLs (Miixr, iMix) still appear in external press and in the sitemap but return 403 to everyone. The brand has undergone a silent product-line rebrand (iMix -> Pro/Sense/Pursuit) that no external knowledge source has caught up with.

Findings

1. Cloudflare 403 blocks on multiple product pages High

   Several product pages, including /products/pursuit, /products/miixr-stealth, and /products/miixr-x7-electric-protein-shake-blender-usa, return 403 to all AI crawlers due to Cloudflare JS challenge/WAF, making them invisible to crawlers despite being listed in the sitemap.

   What to change: Configure Cloudflare to allow AI crawler user agents (GPTBot, ClaudeBot, etc.) through the WAF, or serve static HTML versions of these pages to bots.

2. Cold LLM knowledge stuck on discontinued iMix product line High

   Frontier LLMs describe PROMIXX primarily through the discontinued iMix and iMix Pro products, while the current site has zero mention of 'iMix' and has rebranded to Pro, Sense, and Pursuit. This creates a gap between AI knowledge and actual offerings.

   What to change: Publish a migration guide or announcement page explaining the rebrand, and update external press mentions and backlinks to reflect current product names.

3. Missing FAQPage schema on reviews page Medium

   The reviews page contains FAQ content about usage and cleaning but lacks FAQPage schema markup, reducing its eligibility for rich results in AI-generated answers.

   What to change: Add FAQPage JSON-LD schema to the reviews page for each question-answer pair.

4. Blog articles lack Article schema markup Medium

   Blog pages use Blog/ItemList schema but individual article pages lack Article or NewsArticle markup, limiting their visibility in AI-driven content summaries.

   What to change: Add Article or NewsArticle JSON-LD schema to each blog post page.

5. Missing Review schema for individual testimonials Low

   The reviews page displays testimonials but does not mark them up with Review schema, reducing their potential to appear as rich snippets in AI search results.

   What to change: Add Review schema markup for each individual testimonial on the reviews page.

6. Sitemap contains old product URLs that return 403 Medium

   The sitemap lists URLs for discontinued products (Miixr, iMix) that return 403 to all crawlers, wasting crawl budget and potentially confusing AI agents.

   What to change: Remove discontinued product URLs from the sitemap or redirect them to current product pages.

7. Robots.txt lacks AI-bot-specific rules Low

   The robots.txt uses the standard Shopify template with no specific directives for AI crawlers, missing an opportunity to guide them away from blocked pages.

   What to change: Add explicit disallow rules for AI bots on pages that return 403, or allow them with caveats.

8. External press mentions reference discontinued products Medium

   Press mentions from GQ, T3, and others reference the discontinued Miixr+ Stealth, reinforcing outdated AI knowledge about the product line.

   What to change: Reach out to publishers to update articles with current product names, or create a press page with updated product information.

What's working

• Sophisticated llms.txt and agents.md with UCP and MCP endpoints — The site hosts llms.txt and agents.md documents defining Universal Commerce Protocol versions, MCP tool endpoints, and explicit agent shopping flows, making it technically ready for AI-agent-driven purchasing.
• Rich JSON-LD schema on product pages — Product pages include Product with ProductGroup variants, AggregateRating, gtin13, mpn, sku, priceValidUntil, and BreadcrumbList, providing structured data for AI crawlers.
• Active blog with recent, relevant content — The ACADEMY blog publishes regularly (latest post April 2026) with content on fitness, nutrition, and wellness, providing a strong content marketing engine for AI retrieval.
• Homepage and key product pages accessible to all AI crawlers — The homepage, /products/promixx-pro, and /products/sense return 200 with full content to all 11 tested AI bot user agents.
• Organization and WebSite schema on homepage — The homepage includes Organization and WebSite schema with SearchAction, helping AI crawlers understand the brand and site structure.
• High aggregate ratings on product pages — Product pages show high ratings (Sense: 4.96/23 reviews, Pro: 4.92/600 reviews) in schema, providing strong social proof for AI-generated answers.