AI Site Grade
miro.com — AI Site Grade
Miro's marketing pages return zero visible text to AI crawlers despite permissive access, and the site's sitemap is broken, leaving core product pages undiscoverable.
Miro's AI visibility is undermined by JavaScript-rendered pages that deliver empty shells to crawlers, a broken sitemap, outdated cold knowledge, and missing schema on key pages.
- Findings
- 8
- Evidence checks
- 25
- Completed
- 30 May 2026
Analysis
Miro's AI-Visibility Audit
The site's marketing pages are entirely JavaScript-rendered and return zero words of visible text to any crawler that does not execute JS, yet every major AI bot receives a 200 status with the same empty shell — a critical gap between access permission and actual content delivery.
Crawler Access
All 11 tested AI bots (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Bytespider, Applebot-Extended, Perplexity-User) receive a 200 status on the homepage with identical byte payload (~1 MB) to a browser. No UA-based blocking exists. However, the homepage, /canvas/, and /intelligent-canvas/ all return 0 words of visible text from a plain GET — the content is hydrated client-side by Next.js. The robots.txt contains no AI-bot-specific directives; the catch-all * rule disallows /pricing/pro/, /changelog/, /contact/*, and /templates-v2/, but allows all major marketing paths. The llms.txt returns a 404 (serving the full JS app shell). The sitemap referenced in robots.txt at a230405.sitemaphosting5.com also 404s; the standard /sitemap.xml 404s too. Only the blog at /blog/sitemap.xml is properly indexed with 516+ URLs.
Cold-Knowledge Gap
The LLM's prior knowledge describes Miro as a "digital whiteboard platform" founded in Russia in 2011, rebranded from RealtimeBoard, valued at $17.5B in 2022, with layoffs in 2023 and data-privacy concerns tied to Russian origins. The actual site has completely repositioned — the homepage title is "AI Innovation Workspace" and every page leads with "AI platform for teamwork," "Sidekicks," "Flows," and "Connectors." The site mentions zero about Russian origins, the RealtimeBoard rebranding, or the 2022 valuation. The word "whiteboard" appears only in the footer and URL paths. The site's entire narrative is now about agentic AI collaboration (Canvas 26 in May 2026 launched Sidekicks, Flows, Connectors, Prototypes). The cold model is ~3 years out of date on positioning and has no awareness of the AI product suite.
Schema Posture
The site has a single Organization schema (@id: https://miro.com/#org) with name, URL, logo, and social profiles — present in the <head> of every page. The blog uses proper Article and WebPage schemas with dates, authors, and word counts. However, the core marketing pages (/, /canvas/, /intelligent-canvas/, /ai/ai-overview/) carry zero JSON-LD beyond the Organization block. No SoftwareApplication, WebApplication, Product, or FAQPage schema exists anywhere on the product pages. The /pricing/ page has an FAQ section with 11 questions but uses no FAQPage schema. The blog is the only section with rich schema coverage.
External Signals
The DNS TXT records reveal three OpenAI domain verifications (openai-domain-verification=dv-*), an anthropic-domain-verification, and a cursor-domain-verification — confirming Miro has proactively registered with major AI platforms for crawler trust and potential training data inclusion. The site runs on AWS CloudFront (via x-amz-cf-* headers) with a permissive CSP (default-src 'unsafe-inline' 'unsafe-eval' data: blob: *). External review links point to Capterra, G2, and TrustRadius with "20,000+ reviews" claimed in the footer. The blog is actively publishing (May 2026 posts about Canvas 26), but the main sitemap is broken, meaning Google and other search engines cannot discover the core marketing pages through the declared sitemap path.
Findings
Core marketing pages return zero visible text to crawlers High
The homepage, /canvas/, and /intelligent-canvas/ return 0 words of visible text from a plain GET request. Content is hydrated client-side by Next.js, so AI bots that do not execute JavaScript receive an empty shell despite a 200 status.
What to change: Implement server-side rendering or static generation for key marketing pages to deliver meaningful text content to crawlers.
Primary sitemap returns 404 High
The sitemap referenced in robots.txt at a230405.sitemaphosting5.com returns 404, and the standard /sitemap.xml also returns 404. Only the blog sitemap at /blog/sitemap.xml is accessible, meaning core marketing pages are not discoverable via sitemap.
What to change: Fix the sitemap configuration to serve a valid sitemap at the standard location and update robots.txt accordingly.
llms.txt returns 404 Medium
The /llms.txt endpoint returns a 404, serving the full JS app shell instead of a plain-text file. This prevents AI crawlers from easily discovering the site's content and structure.
What to change: Create a valid llms.txt file listing key URLs and a brief description of the site's content.
LLM cold knowledge is 3 years out of date Medium
The LLM's prior knowledge describes Miro as a digital whiteboard platform with Russian origins and a 2022 valuation, but the site has completely repositioned as an AI Innovation Workspace with no mention of those details. The model has no awareness of the AI product suite (Sidekicks, Flows, Connectors).
What to change: Publish an llms.txt and ensure key product pages contain clear, crawlable text that reinforces the new AI positioning.
Product pages lack SoftwareApplication and FAQPage schema Medium
Core marketing pages carry only an Organization schema. No SoftwareApplication, WebApplication, Product, or FAQPage schema exists on product pages. The /pricing/ page has an FAQ section with 11 questions but no FAQPage schema.
What to change: Add SoftwareApplication schema to product pages and FAQPage schema to the pricing FAQ section.
Multiple key pages return zero words to crawlers High
In addition to the homepage, /canvas/ and /intelligent-canvas/ also return 0 words of visible text. These pages are critical for AI visibility but are invisible to non-JS crawlers.
What to change: Ensure these pages deliver meaningful HTML content server-side.
Robots.txt lacks AI-bot-specific rules Low
The robots.txt has no directives for AI bots like GPTBot or ClaudeBot. While this allows access, it also means no guidance is provided for crawling behavior, and the catch-all rule disallows some paths that may be relevant.
What to change: Consider adding specific directives for AI bots to manage crawl rate and access to key content.
External review signals are not crawlable Low
Web searches for Miro AI reviews returned zero results, indicating that external review sites may not be indexed or are not linking to the site effectively. The site claims 20,000+ reviews on Capterra, G2, and TrustRadius, but these are not surfaced in search.
What to change: Encourage review sites to index reviews and consider adding review schema to the site.
What's working
- All major AI bots receive 200 status on homepage — All 11 tested AI bots receive a 200 status on the homepage with no UA-based blocking, ensuring access permission is not a barrier.
- Organization schema present on all pages — Every page includes an Organization JSON-LD block with name, URL, logo, and social profiles, providing consistent entity identification.
- Blog uses Article and WebPage schema with rich metadata — The blog section uses proper Article and WebPage schemas with dates, authors, and word counts, enhancing discoverability of blog content.
- Domain verified with OpenAI, Anthropic, and Cursor — DNS TXT records show domain verifications for OpenAI, Anthropic, and Cursor, indicating proactive registration with major AI platforms for crawler trust.
- Blog is actively publishing with recent content — The blog has recent posts from May 2026 about Canvas 26, demonstrating ongoing content creation that can be indexed.
- Blog sitemap is properly indexed with 80+ URLs — The blog sitemap at /blog/sitemap.xml returns 200 with 80 URLs, ensuring blog content is discoverable.
- Pricing page contains substantial text content — The /pricing/ page returns 3228 words of visible text, providing detailed information that can be crawled and indexed.
- AI overview page contains 4540 words of text — The /ai/ai-overview/ page returns 4540 words of visible text, offering substantial content about the AI platform.
Track miro.com across AI search
This is one snapshot. Open the interactive report to inspect evidence, or grade another site free.