virtahealth.com — AI Site Grade

Analysis

The $160M Site With Zero Schema

Every page on virtahealth.com — the homepage, the FAQ, the outcomes page, the research page, the blog, the GLP-1 landing page — returns zero JSON-LD schema blocks. Not a single MedicalOrganization, HealthPlanNetwork, FAQPage, Article, Product, or WebSite schema type was detected across any of the eight pages fetched. For a company that claims to reverse type 2 diabetes and cites peer-reviewed studies, this is a structural blind spot that leaves AI crawlers and knowledge panels without structured grounding.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Applebot-Extended, anthropic-ai — receive a 200 with full HTML content identical to the browser baseline (~197KB). The sole exception is Bytespider (TikTok's crawler), which gets a 403 from Cloudflare. The robots.txt is minimal: a single User-agent: * rule disallowing three low-value paths (/health-techology-featured-card-content, /styleguide/, /styleguide) and no AI-bot-specific directives. No llms.txt exists (returns a 404 Webflow error page). The site runs on Webflow behind Cloudflare with HSTS and X-Frame-Options: DENY. The DNS TXT records include an openai-domain-verification token, confirming Virta has proactively verified its domain with OpenAI for potential GPT retrieval integration.

Cold-Knowledge Gap

The LLM's cold knowledge describes Virta as a "digital health company specializing in reversing type 2 diabetes through a nutrition-based intervention" founded in 2014 by Sami Inkinen and Dr. Stephen Phinney. The site itself has evolved significantly beyond this framing. The homepage now leads with "Sustainable Weight Loss and Diabetes Reversal" and prominently features GLP-1 weight loss medications as a core offering — a major strategic pivot the cold model does not capture. The blog's top headline announces "Record and Accelerating Growth at a Scale of More Than $100M in Annualized Revenue" while the newsroom shows $160M in annualized revenue as of September 2025. The cold model knows nothing about the GLP-1 deprescription research, the CMS selection, or the performance guarantee model ("0% GLP-1 Trend").

Content & Schema Posture

The homepage has a clear H1 ("Don't manage metabolic disease. Reverse it.") and strong outcome metrics (31 lbs average weight loss, 63% medication elimination, 1.3% A1c reduction). The FAQ page at /faq contains extensive Q&A content but uses no FAQPage schema. The /diabetes-faq page similarly lacks structured markup. The /research page lists 14+ peer-reviewed publications with study headlines but no ScholarlyArticle or MedicalStudy schema. The /outcomes page presents clinical results with footnoted citations but no structured data. The blog has dozens of articles but no BlogPosting or Article schema. The site uses no breadcrumb, no sitelinks search box, no Organization schema on any page.

External Signals

The newsroom shows press mentions in Men's Health, Fox News, Real Simple, CFO Thought Leader, and podcast appearances. The DNS records reveal integrations with Stripe, Salesforce, HubSpot, Zendesk, Firebase, Google Workspace, SendGrid, and Mailchimp — a mature enterprise tech stack. The anthropic-domain-verification TXT record confirms Virta has also verified with Anthropic for Claude-powered use cases. No recent Reddit threads or negative press surfaced in searches. The Wayback Machine returned no snapshot, suggesting the site may block archiving or is relatively recent in its current form.

Findings

1. Zero JSON-LD schema on all pages High

   Every page on virtahealth.com returns no JSON-LD schema blocks. No MedicalOrganization, FAQPage, Article, or WebSite schema was detected across eight pages, leaving AI crawlers without structured grounding.

   What to change: Add JSON-LD schema for Organization, WebSite, FAQPage, Article, and MedicalOrganization on relevant pages.

2. Cold knowledge misses GLP-1 weight loss pivot High

   The LLM's cold knowledge describes Virta as a diabetes reversal company, but the site now leads with GLP-1 weight loss medications as a core offering. This gap means AI assistants may provide outdated information.

   What to change: Publish an llms.txt file and ensure key pages are indexed with updated schema to reflect the GLP-1 offering.

3. No llms.txt file available Medium

   The site returns a 404 for llms.txt, missing an opportunity to guide AI crawlers with a curated summary of key pages and content.

   What to change: Create an llms.txt file listing important pages like /research, /outcomes, /faq, and /blog.

4. FAQ pages lack FAQPage schema Medium

   The /faq and /diabetes-faq pages contain extensive Q&A content but use no FAQPage structured data, reducing their visibility in AI-generated answers.

   What to change: Add FAQPage schema to both FAQ pages.

5. Blog posts lack Article schema Medium

   The blog page lists articles but no BlogPosting or Article schema is present, limiting structured representation in search and AI results.

   What to change: Add Article or BlogPosting schema to each blog post.

6. Research page lacks ScholarlyArticle schema Medium

   The /research page lists 14+ peer-reviewed publications but uses no ScholarlyArticle or MedicalStudy schema, missing a chance to surface studies in AI knowledge panels.

   What to change: Add ScholarlyArticle schema to each study entry on the research page.

7. No Organization schema on any page Medium

   No page includes Organization or MedicalOrganization schema, which would help AI crawlers understand the entity behind the site.

   What to change: Add Organization schema with name, logo, and description to the homepage and key pages.

8. No breadcrumb schema on any page Low

   The site does not implement BreadcrumbList schema, which helps crawlers understand site hierarchy and can enhance search snippets.

   What to change: Add BreadcrumbList schema to all pages.

9. Bytespider crawler blocked by Cloudflare Low

   Bytespider (TikTok's crawler) receives a 403 from Cloudflare, while all other major AI crawlers are allowed. This may limit visibility on TikTok search.

   What to change: Consider allowing Bytespider if TikTok search is a target channel.

10. Site blocks Wayback Machine archiving Low

    The Wayback Machine returned no snapshot, suggesting the site may block archiving, which can reduce historical context for AI training data.

    What to change: Allow archiving by removing any blocking directives in robots.txt or headers.

What's working

• All major AI crawlers allowed full access — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and others receive 200 with full HTML content, ensuring AI crawlers can index the site's content.
• Domain verified with OpenAI and Anthropic — DNS TXT records include openai-domain-verification and anthropic-domain-verification tokens, enabling potential GPT and Claude retrieval integrations.
• Homepage has clear H1 and outcome metrics — The homepage features a strong H1 ('Don't manage metabolic disease. Reverse it.') and key metrics (31 lbs weight loss, 63% medication elimination), which are easily parseable by crawlers.
• Mature enterprise tech stack with Stripe, Salesforce, HubSpot — DNS records show integrations with Stripe, Salesforce, HubSpot, Zendesk, Firebase, and others, indicating a robust infrastructure.
• Newsroom lists press mentions in major outlets — The newsroom page shows press mentions in Men's Health, Fox News, Real Simple, and CFO Thought Leader, providing external credibility signals.
• Robots.txt is minimal and does not block AI bots — The robots.txt only disallows three low-value paths and has no AI-bot-specific directives, ensuring broad crawler access.