Security

Learning data deserves careful, layered protection.

Archie Math combines application, infrastructure, access-control, and research-governance safeguards. This page describes our current approach; it is not a claim of external security certification.

Identity and access

Authenticated tutoring endpoints, password hashing, server-side sessions, CSRF protection for state-changing account actions, and role-based separation between learner and educator analytics.

Transport and browser protection

HTTPS in production, HTTP Strict Transport Security, content-type protections, restricted browser permissions, and a privacy-preserving referrer policy.

Application safeguards

Input validation, output escaping and sanitisation, rate and credit controls, server-side authorization checks, release tests, health checks, and operational logging.

Research governance

Research data capture is separated from ordinary product use through approved project, cohort, consent, and participation workflows. Provisional model estimates remain non-decisioning until validated.

Responsible disclosure

If you believe you have found a security or privacy issue, please do not access, alter, or download data that is not yours. Send a concise report with the affected URL, reproduction steps, impact, and your contact details to help@archiemath.com.

Please allow reasonable time for investigation before public disclosure. Do not perform denial-of-service testing, automated credential attacks, social engineering, or testing that disrupts learners or educators.

Data and model boundaries

Archie Math’s AI responses may use approved model providers. Authentication and authorization remain enforced by Archie Math rather than delegated to the model. Knowledge-tracing outputs are treated as research signals—not grades, diagnoses, or automated academic decisions—unless and until an institution approves a separately validated use.

Questions

For account, privacy, or security support, contact help@archiemath.com. See also our Privacy Policy, Terms, and Consent information.