amiralearning.com — AI Site Grade

Analysis

The llms.txt Is an RTF File, Not Plain Text

The site at amiralearning.com serves its llms.txt as a Rich Text Format (RTF) document (content-type text/plain but containing {\rtf1\ansi... headers and font-table markup), not as a clean Markdown file. An AI crawler reading this file will parse raw RTF control codes instead of the intended section headings and link references, effectively nullifying the guidance the file was designed to provide.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, Google-Extended, PerplexityBot, ChatGPT-User, OAI-SearchBot, Applebot-Extended, and anthropic-ai — receive a 200 status with the full HTML page (205 KB) on the homepage. Only Bytespider is blocked (403). The robots.txt contains a single line (Sitemap: https://amiralearning.com/sitemap.xml) with zero user-agent rules and no AI-bot directives. The site is hosted on Cloudflare (AWS DNS) and returns server-rendered HTML with ~846 words of visible text on the homepage — no JS-shell risk. The sitemap indexes 260 URLs.

Cold-Knowledge Gap

The LLM model queried about "Amira Learning" returned a factually incorrect claim: that the company was "acquired by HMH (Houghton Mifflin Harcourt) in 2023." The site itself makes no mention of any acquisition. The /company page names Mark Angel as Founder & CEO and describes Amira as an independent entity spun from Carnegie Mellon research. The model also stated the company serves "primarily K-5 students" — the site consistently says PreK-8. The model described the product as a "reading tutor" — the site now positions it as a "Learning Agent" connecting assessment, instruction, and tutoring into one suite (Amira ISIP Assess, Amira Instruct, Amira Tutor). The model's knowledge is stale by at least one major product rebranding cycle.

Schema Posture

The homepage carries Organization and WebSite JSON-LD with complete sameAs links, founding date (2018), contact info, and multilingual support. The /amira-reading-suite page adds a SoftwareApplication schema. However, the /research page and /blog page contain zero JSON-LD — no ScholarlyArticle, Article, or BreadcrumbList schema on pages where AI engines would most benefit from structured metadata about studies and thought-leadership content.

External Signals

The site references Istation heavily — including a dedicated /istation-to-amira migration page — suggesting Amira either acquired Istation's user base or rebranded from it. The DNS includes an openai-domain-verification token, confirming proactive OpenAI relationship management. The site claims 5.5M+ students, 4K+ districts, and 11B+ words read, but web search returned zero external results for the brand name, making independent verification of these claims impossible through the search tool used.

Findings

1. llms.txt served as RTF file, not plain text High

   The llms.txt file contains RTF control codes ({\rtf1\ansi...) instead of clean Markdown, making it unparseable by AI crawlers.

   What to change: Replace the llms.txt file with a plain text or Markdown file containing the intended section headings and link references.

2. LLM knowledge about Amira Learning is stale and factually incorrect High

   The LLM model returned incorrect claims: that Amira was acquired by HMH in 2023, that it serves K-5 (site says PreK-8), and that it is a reading tutor (site now positions as a Learning Agent). The site shows no acquisition evidence.

   What to change: Publish a clear, authoritative page detailing the company's history, product evolution, and current positioning to correct LLM misconceptions.

3. Research and blog pages lack JSON-LD schema Medium

   The /research and /blog pages contain no structured data (ScholarlyArticle, Article, BreadcrumbList), missing opportunities for AI crawlers to understand and cite this content.

   What to change: Add appropriate JSON-LD schema (e.g., ScholarlyArticle, Article, BreadcrumbList) to the research and blog pages.

4. No external search results found for the brand High

   Multiple web searches for 'Amira Learning' and related terms returned zero results, indicating extremely low external visibility and making independent verification of claims impossible.

   What to change: Investigate and improve search engine indexing and off-site presence (press, backlinks, social signals) to increase external visibility.

5. Robots.txt has no AI crawler directives Low

   The robots.txt file contains only a sitemap directive and no user-agent rules, leaving AI crawler access entirely to default behavior.

   What to change: Add explicit allow/disallow rules for AI crawlers to ensure intended access and prevent accidental blocking.

What's working

• All major AI crawlers receive 200 status with full HTML — GPTBot, ClaudeBot, Google-Extended, and others get the full server-rendered HTML homepage (846 words) with no JS-shell issues.
• Homepage has complete Organization and WebSite JSON-LD — The homepage includes Organization and WebSite schema with sameAs links, founding date, contact info, and multilingual support.
• Product page includes SoftwareApplication schema — The /amira-reading-suite page adds SoftwareApplication JSON-LD, helping AI crawlers understand the product offering.
• DNS includes openai-domain-verification token — The site has proactively set up an OpenAI domain verification token, indicating active relationship management with OpenAI.
• Sitemap indexes 260 URLs — The sitemap contains 260 URLs, providing good coverage for crawlers to discover site content.