releaf.co.uk — AI Site Grade

Analysis

Releaf's llms.txt is more authoritative than what AI models actually know about the brand

The site operates an unusually mature AI-visibility setup — including a detailed llms.txt, an OpenAI domain verification TXT record, and zero bot blocking — yet the cold LLM knowledge about Releaf is stale and partially inaccurate, creating a gap between what the site claims and what AI engines recall.

Crawler Access

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receives a full 200 response with identical content to a browser baseline. The site runs on Vercel behind Cloudflare DNS, with no UA-based blocking, no JS-shell risk, and no WAF interference. The robots.txt contains no AI-bot-specific rules; the wildcard * rule disallows only thin pages (/education/article/*, /campaign/*, /api/) that are unlikely to be missed. An openai-domain-verification TXT record confirms proactive OpenAI relationship management.

Cold-Knowledge Gap

The cold LLM describes Releaf as having a "Cannabis Care Plan" subscription (the actual name is Releaf+), cites a Trustpilot rating of "3.5–4 stars" (the site shows 4.8/5 from 7,000+ reviews), and mentions "delays in medication delivery" as a reputational signal — a claim contradicted by the 4.8 rating and the 93% of reviews in the top two bands. The model also references "Grow Pharma" as a supplier, whereas the site prominently names Glass Pharms and SOMAÍ as direct partners. The model is unaware of Releaf Protect (24/7 legal support via Irwin Mitchell), the Releaf+ rewards programme (1,000+ brand discounts), and the claim of being the UK's largest private medical cannabis clinic with 25,000+ active patients.

Schema Posture

The site deploys rich JSON-LD across all key pages: MedicalOrganization with full contact points, WebSite with SearchAction, WebPage with AboutPage and CollectionPage subtypes, Service with UnitPriceSpecification on the pricing page, Product with AggregateOffer for cannabis flower and oil, FAQPage with individual Question/Answer entities on FAQ pages, Article on blog posts, VideoObject on the homepage, and BreadcrumbList throughout. The MedicalOrganization schema includes additionalType linking to Wikidata entries and a sameAs array covering Facebook, Instagram, LinkedIn, Twitter, Wikipedia, and CQC. This is a near-complete schema implementation — the only missing element is MedicalWebPage or MedicalCondition types on condition-specific pages, which would strengthen disease-specific retrieval.

External Signals

The site links to Trustpilot with a displayed 4.8/5 rating and embeds dozens of patient testimonials. The blog actively publishes comparison content (e.g., "most trusted medical cannabis clinic UK") with tables benchmarking against competitors like Alternaleaf, Mamedica, and Curaleaf. External links point to releaf.com (a separate international site) and releaf.com/de. No negative press or regulatory sanctions surfaced in searches. The /faq path returns a 404 — the actual FAQ hub lives at /faqs/general-faqs and similar sub-paths, a minor navigation inconsistency.

Findings

1. Cold LLM knowledge about Releaf is stale and partially inaccurate High

   The cold LLM describes Releaf's subscription as 'Cannabis Care Plan' (actual name is Releaf+), cites a Trustpilot rating of 3.5–4 stars (site shows 4.8/5), and mentions 'delays in medication delivery' as a reputational signal, contradicting the 4.8 rating. The model is unaware of Releaf Protect, Releaf+ rewards, and the claim of being the UK's largest private medical cannabis clinic with 25,000+ active patients.

   What to change: Ensure that key brand facts (Releaf+, Trustpilot rating, patient count, partnerships) are prominently featured in structured data and on the homepage, and consider submitting an updated llms.txt or using OpenAI's domain verification to improve model recall.

2. FAQ path returns 404, actual FAQ hub at different URL Medium

   The URL /faq returns a 404 error, while the actual FAQ content is located at /faqs/general-faqs and similar sub-paths. This creates a broken navigation path for users and crawlers.

   What to change: Redirect /faq to /faqs/general-faqs or serve the FAQ hub at /faq.

3. Missing MedicalWebPage and MedicalCondition schema on condition pages Medium

   Condition-specific pages (e.g., /conditions) lack MedicalWebPage or MedicalCondition schema types, which would strengthen disease-specific retrieval in AI and search results.

   What to change: Add MedicalWebPage and MedicalCondition schema to condition pages, including relevant symptoms and treatment details.

4. No external review pages (Trustpilot, Reddit) indexed in search results Medium

   Web searches for Releaf on Trustpilot and Reddit returned zero results, indicating that external review platforms may not be indexed or linked effectively, reducing social proof signals.

   What to change: Encourage patients to leave reviews on Trustpilot and Reddit, and link to those profiles from the site to improve indexing.

5. Robots.txt disallows thin pages that may include useful content Low

   The robots.txt disallows /education/article/* and /campaign/*, which may contain valuable educational content that AI crawlers cannot access.

   What to change: Review the disallowed paths and allow AI crawlers to access educational articles if they are not thin or duplicate content.

What's working

• Detailed llms.txt published with comprehensive brand information — Releaf publishes a 26KB llms.txt file that provides authoritative brand information, helping AI models access accurate data about the clinic.
• OpenAI domain verification TXT record present — An openai-domain-verification TXT record indicates proactive management of AI relationships, allowing OpenAI to verify the domain.
• All major AI crawlers receive full access with identical content — Every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, etc.) receives a 200 response with content identical to a browser baseline, with no UA-based blocking or JS-shell risk.
• Near-complete schema implementation across key pages — The site deploys rich JSON-LD including MedicalOrganization, WebSite, Service with pricing, Product with offers, FAQPage, Article, VideoObject, and BreadcrumbList, covering most entity types.
• Trustpilot rating of 4.8/5 displayed on site with patient testimonials — The site prominently displays a Trustpilot rating of 4.8/5 from 7,000+ reviews and embeds patient testimonials, providing strong social proof.
• Blog publishes comparison content benchmarking against competitors — The blog includes articles like 'most trusted medical cannabis clinic UK' with tables comparing Releaf to Alternaleaf, Mamedica, and Curaleaf, which can attract AI attention for competitive queries.
• Releaf Protect and Releaf+ programmes detailed on site — Pages for Releaf Protect (24/7 legal support via Irwin Mitchell) and Releaf+ (rewards programme) provide unique value propositions that differentiate the clinic.
• FAQ hub at /faqs/general-faqs with FAQPage schema — The FAQ page contains 4,576 words of content with FAQPage JSON-LD, providing structured answers to common questions that AI can use for featured snippets.