onepeloton.com — AI Site Grade

Analysis

Peloton's AI crawlers get full access but find almost nothing structured

All 11 AI bot user-agents (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, etc.) receive a 200 status with full HTML content from the Vercel-hosted homepage — no blocks, no JS shells, no UA-based discrimination. Yet the site has zero JSON-LD schema on the homepage, no sitemap.xml (returns 404), no llms.txt (returns 404), and a robots.txt that mentions zero AI crawlers by name — only generic Disallow rules for /delivery/, /orderhistory/, /mymembership/, and /calculator/results/.

Crawler Access

Every AI bot tested returns identical byte-size content (530,682 bytes) matching the browser baseline. The site runs on Vercel behind Cloudflare DNS, with strong security headers (HSTS preload, CSP with frame-ancestors). No bot is blocked, rate-limited, or served a thin shell. However, the absence of a sitemap.xml means AI crawlers must discover Peloton's ~25+ significant pages (product pages, blog, classes, offers) purely through internal links — no crawl budget optimization, no priority signals, no lastmod hints. The robots.txt at 116 bytes is the thinnest possible configuration for a site of this scale.

Cold-Knowledge Gap

The LLM cold-knowledge snapshot of Peloton contains stale pricing (Bike+ quoted at $2,995 vs. the site's actual $2,695) and references the Tread+ recall from 2021 and CEO Barry McCarthy stepping down in 2024 — events the site itself never addresses. The model knows Peloton as a post-pandemic cautionary tale (layoffs, stock decline) but has no awareness of Peloton IQ, the company's major AI-powered feature suite launched October 2025. The site heavily promotes Peloton IQ across every product page, yet the cold model has zero knowledge of it. This is the single largest gap: Peloton is actively building an AI narrative but AI engines don't know it exists.

Schema Posture

The homepage and most key pages carry no structured data at all. The /bike-plus and /tread-plus product pages each have a single WebPage wrapper containing a minimal Product entity with only name, brand, and a BuyAction target — no offers (price, currency, availability), no aggregateRating, no review, no description. The /shop/bike-plus and /shop/bike checkout pages have zero schema. The blog uses Article schema but with a dateModified of 1970-01-01T00:00:00.000Z — a clear data-pipeline bug. The /classes page, /blog index, and /peloton-iq page all have no schema at all. For a site selling hardware at $1,695–$6,695, the absence of rich Product schema with pricing is a material gap for AI-driven shopping answers.

External Signals

The DNS TXT records reveal an unusually deep vendor stack: 15+ Google site verifications, multiple Stripe verifications, OneTrust, Segment, Optimizely, New Relic, and integrations with Drift, Calendly, Zendesk, Salesforce, Shopify, and Amazon SES. This signals a mature but fragmented tech ecosystem — the kind where structured data often falls through the cracks between teams. The blog ("The Output") is content-rich with 570+ words on the index page and long-form articles, but none of this content is surfaced through sitemaps or schema-linked to product pages.

Findings

1. Sitemap.xml returns 404, crippling AI crawl discovery High

   The sitemap.xml at onepeloton.com returns a 404 error, preventing AI crawlers from efficiently discovering the site's 25+ significant pages. No crawl budget optimization, priority signals, or lastmod hints are provided.

   What to change: Generate and serve a valid sitemap.xml listing all key pages (product, blog, classes, offers) with appropriate priority and lastmod values.

2. llms.txt file missing, no AI guidance provided Medium

   The llms.txt file returns a 404, meaning Peloton provides no curated guidance for AI crawlers about which pages to prioritize or how to interpret content.

   What to change: Create an llms.txt file that lists key pages and provides brief descriptions for AI crawlers.

3. Robots.txt does not name any AI crawler user-agents Low

   The robots.txt file is minimal (116 bytes) and only contains generic Disallow rules for non-content paths. No AI-specific user-agents (GPTBot, ClaudeBot, etc.) are mentioned, leaving their access entirely to default behavior.

   What to change: Add explicit rules for major AI crawlers to ensure they are allowed and to signal crawl preferences.

4. Homepage has zero JSON-LD structured data High

   The homepage contains no JSON-LD schema of any type, missing opportunities to provide entity context (Organization, WebSite, Product) to AI crawlers.

   What to change: Add JSON-LD structured data for Organization, WebSite, and relevant Product or ItemList schemas to the homepage.

5. Product pages have minimal schema missing price and offers High

   The Bike+ and Tread+ product pages include only a basic Product entity with name and brand, but lack offers (price, currency, availability), aggregateRating, review, and description. The /shop/ pages have no schema at all.

   What to change: Add complete Product schema with offers (price, currency, availability), aggregateRating, and description to all product and shop pages.

6. Blog Article schema has invalid dateModified of 1970 Medium

   The blog article on Peloton IQ uses Article schema but with a dateModified value of 1970-01-01T00:00:00.000Z, indicating a data-pipeline bug that could confuse crawlers and degrade trust.

   What to change: Fix the dateModified field in Article schema to reflect the actual last-modified date of the article.

7. Classes page lacks any structured data Medium

   The /classes page, a key content hub, has no JSON-LD schema, missing the chance to mark up classes as Events or CreativeWork.

   What to change: Add appropriate schema (e.g., Event, CreativeWork) to the classes page to help AI crawlers understand the content.

8. Peloton IQ page has no structured data High

   The /peloton-iq page, which describes the company's major AI feature, has no schema at all, making it invisible to AI engines looking for AI-related content.

   What to change: Add JSON-LD schema (e.g., SoftwareApplication, Product) to the Peloton IQ page to describe the AI feature set.

9. LLM cold knowledge contains stale pricing and outdated recall info High

   The LLM knowledge snapshot quotes Bike+ at $2,995 (actual site price $2,695) and references the 2021 Tread+ recall and 2024 CEO departure, which the site never addresses. This outdated information can lead to fabricated facts in AI answers.

   What to change: Publish structured data with current pricing and proactively address historical issues on the site to correct the knowledge gap.

10. Peloton IQ is completely unknown to LLMs despite heavy site promotion High

    The cold LLM knowledge has zero awareness of Peloton IQ, the company's AI-powered feature suite launched October 2025, even though the site promotes it across all product pages. This is the largest gap: Peloton's AI narrative is invisible to AI engines.

    What to change: Add structured data to the Peloton IQ page and blog, and consider submitting to AI crawler indexes or using llms.txt to highlight this content.

11. Blog index page lacks structured data Low

    The /blog index page has no schema, missing the opportunity to mark up the list of articles as a CollectionPage or BlogPosting.

    What to change: Add CollectionPage or BlogPosting schema to the blog index page.

12. Fragmented tech ecosystem risks structured data ownership gaps Medium

    DNS TXT records reveal 15+ Google site verifications and integrations with Stripe, OneTrust, Segment, Optimizely, New Relic, Drift, Calendly, Zendesk, Salesforce, Shopify, and Amazon SES. This mature but fragmented stack often leads to structured data falling through the cracks between teams.

    What to change: Establish a centralized structured data governance process to ensure consistent schema implementation across all teams and vendors.

What's working

• All 11 AI bots receive full HTML content with no blocks — Every tested AI bot (GPTBot, ClaudeBot, PerplexityBot, etc.) gets a 200 status with full HTML content from the homepage, matching the browser baseline. No bot is blocked, rate-limited, or served a thin shell.
• Strong security headers with HSTS and CSP — The site uses HSTS preload and Content-Security-Policy with frame-ancestors, providing robust security that also signals trustworthiness to AI crawlers.
• Blog 'The Output' is content-rich with long-form articles — The blog index page contains 570+ words and individual articles are long-form, providing substantial text content that AI crawlers can index for knowledge.
• Product pages include minimal Product schema with brand and name — The Bike+ and Tread+ product pages have a basic Product entity with name and brand, providing a foundation that can be expanded.
• Peloton IQ is prominently featured across product pages — The site heavily promotes Peloton IQ on every product page and dedicated page, signaling its importance to human visitors and providing content that can be structured for AI.
• FAQ section present on Bike+ page with potential for FAQ schema — The Bike+ page includes an FAQ section (#faq) that could be marked up with FAQPage schema to enable rich results.
• Vercel hosting provides fast, reliable delivery — The site is hosted on Vercel, which offers edge caching and fast global delivery, ensuring AI crawlers get quick responses.
• No JavaScript shell served to AI crawlers — All AI bots receive the same full HTML content as browsers, meaning no critical content is hidden behind JavaScript execution.