goteleport.com — AI Site Grade

Analysis

Teleport (goteleport.com) — AI-Visibility Audit

The site has fully repositioned as "The AI Infrastructure Identity Company" with extensive AI-agent security content, yet frontier LLMs queried cold still describe Teleport as an "identity-aware proxy" company whose AI angle is a natural-language query tool called "Teleport Assist" — a product that no longer appears anywhere on the site.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, anthropic-ai, Bytespider, Applebot-Extended — receive a 200 with identical full HTML content as a browser. No UA-based blocking exists. The site runs on Cloudflare + AWS CloudFront with Next.js server-side rendering, so no JS-rendering risk for crawlers. The robots.txt has a single wildcard rule (User-agent: *) that disallows only utility paths (/teleport.sh/, /teleconsole/, /gravity/, /categories/, /_shared/, /sandbox/, /404/, /blog/404/, /docs/404/) and explicitly allows /api/og/*. No AI-bot-specific directives exist — a missed opportunity to signal crawl priority to GPTBot or ClaudeBot for the new AI content. The llms.txt is present, comprehensive (16KB), and well-structured with platform architecture, protected resources, key terms, and external links — a strong signal.

Cold-Knowledge Gap

The cold LLM describes Teleport as an "identity-aware proxy" company founded in 2015, focused on SSH key management and VPN replacement, with "Teleport Assist" as its AI feature. The actual site has completely erased that positioning. The homepage tagline is "Unified Identity Securing Classic & AI Infrastructure." The brand now calls itself "The AI Infrastructure Identity Company." Products include the Agentic Identity Framework, Beams (trusted agent runtimes in Firecracker VMs), MCP security, and AI Session Summaries. The word "proxy" does not appear in the homepage H1 or meta description. The cold model's prior is 12–18 months stale — it references a 2024 security incident and a funding round that are not mentioned on the site. The model knows nothing about the Agentic Identity Framework, Beams, or the AI infrastructure identity narrative that now dominates the homepage.

Schema Posture

The homepage carries Corporation and WebSite schema with SearchAction — correct but minimal. The AI infrastructure page (/platform/ai-infrastructure/) and the agentic AI use-case page (/use-cases/agentic-ai/) both include FAQPage schema with substantive Q&A about MCP security, agent identity, and model protection — a strong answer-format signal. The blog post on the Agentic Identity Framework uses BlogPosting schema with author and date. However, the homepage itself has no FAQPage, no Product schema, and no SoftwareApplication schema. The Corporation schema lists alternateName: "Gravitational Inc." — the legal entity name, which is not how the brand presents itself anywhere on the site.

Content & Signals

The site has deep AI content — dedicated platform pages for AI infrastructure, agentic identity framework, MCP security, and Beams; a blog with dozens of AI-tagged posts; FAQ sections on every key page. The blog sitemap contains hundreds of posts with lastmod dates as recent as May 2026. The docs site (Docusaurus, 375KB+ served to GPTBot) is fully crawlable. The pricing page includes inline FAQ about licensing models. External signals include G2 and Gartner review links, a 20K-star GitHub repo, and customer logos from Nasdaq, IBM, Doordash, and Elastic. DNS TXT records show domain verification for OpenAI (openai-domain-verification), Anthropic (anthropic-domain-verification), and Tailscale — confirming active partnerships with the two leading AI companies.

Surprising Findings

The sitemap.xml at the root returns 404 — the actual sitemap index lives at /sitemapindex.xml (referenced in robots.txt). The list_known_urls tool found only 25 URLs from the homepage crawl, while the real sitemap index contains 1,894+ URLs across three sub-sitemaps (/next_sitemap.xml, /blog/sitemap.xml, /docs/sitemap.xml). This means any AI crawler or discovery tool that naively checks /sitemap.xml gets nothing. The robots.txt helpfully includes a comment pointing to /llms.txt — a rare and commendable practice. The site uses s-maxage=31536000 (1-year cache) on CloudFront, meaning stale content could persist in CDN caches for crawlers that respect edge cache headers.

Findings

1. Frontier LLMs describe Teleport as an 'identity-aware proxy' with a defunct AI product High

   Cold queries to GPT-4 and Claude return a 12–18 month stale description of Teleport as an 'identity-aware proxy' company focused on SSH key management and VPN replacement, with 'Teleport Assist' as its AI feature. The actual site has completely erased that positioning and now brands itself as 'The AI Infrastructure Identity Company' with products like Agentic Identity Framework, Beams, and MCP security.

   What to change: Submit updated brand and product information to LLM training data sources (e.g., Wikipedia, Crunchbase, press releases) and ensure the homepage and key pages use clear, consistent structured data that reinforces the new positioning.

2. Root sitemap.xml returns 404, blocking naive crawler discovery High

   The standard sitemap URL at /sitemap.xml returns a 404 error. The actual sitemap index lives at /sitemapindex.xml and contains 1,894+ URLs across three sub-sitemaps. Any AI crawler or discovery tool that naively checks /sitemap.xml gets nothing, severely limiting content discovery.

   What to change: Create a redirect from /sitemap.xml to /sitemapindex.xml, or serve the sitemap index at the standard location.

3. Homepage lacks FAQPage, Product, and SoftwareApplication schema Medium

   The homepage carries only Corporation and WebSite schema with SearchAction. It has no FAQPage, Product, or SoftwareApplication schema to help LLMs understand the product offerings and answer common questions directly.

   What to change: Add FAQPage schema with Q&A about the company's AI infrastructure identity focus, and add Product or SoftwareApplication schema for the core platform.

4. Corporation schema lists legal entity name 'Gravitational Inc.' instead of brand name Medium

   The homepage's Corporation schema uses alternateName: 'Gravitational Inc.', which is the legal entity name and not how the brand presents itself anywhere on the site. This could confuse LLMs that extract entity information from schema.

   What to change: Update the Corporation schema to use 'Teleport' as the primary name and remove or de-emphasize 'Gravitational Inc.' unless legally required.

5. Robots.txt lacks AI-bot-specific crawl priority directives Low

   The robots.txt has a single wildcard rule and does not set crawl-delay or allow/disallow directives for specific AI bots like GPTBot or ClaudeBot. This is a missed opportunity to signal which AI-relevant content should be prioritized.

   What to change: Add explicit allow rules for GPTBot and ClaudeBot to key AI content paths (e.g., /platform/ai-infrastructure/, /use-cases/agentic-ai/, /blog/ai-*) to signal crawl priority.

6. CloudFront cache TTL of 1 year may serve stale content to crawlers Low

   The site uses s-maxage=31536000 (1 year) on CloudFront, meaning stale content could persist in CDN caches for crawlers that respect edge cache headers, potentially serving outdated information.

   What to change: Reduce the CDN cache TTL for HTML content to a shorter duration (e.g., 1 hour) to ensure crawlers receive fresh content.

What's working

• Comprehensive llms.txt file (16KB) with platform architecture and key terms — The site provides a well-structured llms.txt file that includes platform architecture, protected resources, key terms, and external links, giving LLMs a strong starting point for understanding the product.
• FAQPage schema on AI infrastructure and agentic AI pages with substantive Q&A — The /platform/ai-infrastructure/ and /use-cases/agentic-ai/ pages include FAQPage schema with detailed Q&A about MCP security, agent identity, and model protection, which helps LLMs extract structured answers.
• Deep AI content with dedicated platform pages and blog posts — The site has extensive AI-focused content including dedicated pages for AI infrastructure, agentic identity framework, MCP security, and Beams, plus a blog with dozens of AI-tagged posts and recent lastmod dates.
• DNS verification records for OpenAI and Anthropic confirm active partnerships — DNS TXT records include openai-domain-verification and anthropic-domain-verification, confirming that Teleport has active partnerships with the two leading AI companies, which can boost credibility with their crawlers.
• All major AI crawlers receive full server-rendered HTML with no blocking — GPTBot, ClaudeBot, PerplexityBot, and other AI crawlers all receive a 200 status with identical full HTML content as a browser, with no UA-based blocking or JS-rendering issues.
• Robots.txt includes a comment pointing to llms.txt — The robots.txt file helpfully includes a comment directing crawlers to the llms.txt file, a rare and commendable practice that aids AI discovery.
• Pricing page includes inline FAQ about licensing models — The pricing page contains an FAQ section that answers common questions about licensing, which can be extracted by LLMs for accurate pricing information.