Methodology

How to use this directory

• Read the category paragraph before the category table. The paragraph states what the category is for, what it gets wrong, and what human oversight it requires; the table is a comparison instrument, not a shopping list.
• Use the category summary for a high-level view of the market, of the provenance of each group of records, and of the oversight burden each category carries.
• Use the tool tables to compare vendor, principal use, modality, deployment model, and API availability — and read the NI Tier as a design requirement for the workflow you will build around the tool.
• Follow the official URL and verification source before procurement, citation, teaching, or any high-stakes use. Record the date on which you verified.
• Treat category placement as a primary-use classification. Many platforms span several domains, and a tool used outside its listed category may carry a higher oversight burden than its tier indicates.
• Apply the educator and family checklist in Appendix B before introducing any tool to learners, and the eleven-domain crosswalk in Appendix C when aligning adoption to institutional or curricular standards.

How the tier is assigned

The tier is a pedagogical and ethical classification made by Doctor Vermeille Global EdTech & Research LLC. It is not a vendor claim and not a safety audit.

These tiers are the firm's professional judgment, not an audit finding.

Framework crosswalk

Caveats and limits

• Temporal drift: product names, ownership, features, pricing, deployment options, language support, and API access change rapidly. Records reflect the verification dates shown and require re-verification before reliance.
• Coverage selectivity: this is a curated directory, not a claim to enumerate every AI product worldwide. Its coverage tilts toward English-language, commercially marketed, United States and European products, which is a limitation of the market's visibility and of the compilation method alike.
• Category ambiguity: tools may reasonably fit more than one category; placement reflects their principal documented use, and a tool used outside that use may carry different risks and a higher oversight burden.
• Vendor evidence: official product pages establish product identity and marketed capabilities. They do not establish effectiveness, accuracy, fairness, accessibility, pedagogical value, or fitness for any particular learner or institution.
• Tier subjectivity: NI Tier assignments are the professional judgment of Doctor Vermeille Global EdTech & Research LLC applied to each tool's documented principal use. They are conservative, contestable, and offered as a reasoned starting point for institutional review rather than as a finding of fact.
• Structural error: generative systems produce confident, fluent output that may be incomplete, inaccurate, biased, or fabricated. This is a property of how the systems work, not a defect that careful vendor selection removes. Verification is required at every tier.
• Human oversight: qualified human judgment remains primary, and remains legally and morally accountable, in healthcare, finance, law, education, employment, security, and every other domain in which people are affected by what a system produces.
• No magic: nothing in this directory is an oracle, a mind, a colleague, or a friend. Every entry is a manufactured system built by a company from data gathered under particular conditions, optimized for particular objectives, and sold for particular reasons. Reading it that way is the beginning of AI literacy.

Sources reviewed