Anvaya AI / Security overview

Security overview

Last updated: 26 October 2026

How Anvaya AI protects your organization's knowledge. For a detailed security questionnaire or a private deployment, contact us.

Isolation and access control

  • Each organization's documents, users and usage are isolated.
  • Roles (owner, admin, member, viewer), groups and group-restricted folders.
  • Answers and suggestions only draw on documents the user may open.
  • Periodic access reviews for administrators.

Authentication

  • Short-lived access tokens with server-side session revocation.
  • Login lockout after repeated failures; password changes revoke other sessions.
  • One-time, expiring tokens for invitations and password resets.

Documents

  • File type and content signature checks; malware scanning before indexing.
  • Integrity hashes and storage reconciliation.
  • Legal hold and retention controls.

Infrastructure

  • HTTPS everywhere with HSTS and a strict content security policy.
  • Firewall limited to web and administrative access; services bound to the local host.
  • Nightly encrypted backups with restore testing; automatic security updates.

AI processing

Search embeddings and the Nepali language engine run on our own server. To write an answer, only the passages relevant to the question are sent to Anthropic's Claude API; under Anthropic's commercial terms that data is not used for model training.

Report a vulnerability

Found a security issue? Email info@matrixnet.com.np with “Security report” in the subject. Please don't access other people's data or disrupt the service while testing.