Product

Security and trust

Effective date: 1 September 2026.

This page describes how the platform protects data and how responsibility around AI is designed. We apply the practices listed below.

Contents

  1. How we protect data
  2. AI and human decisions
  3. Hosting and sub-processors
  4. Contact

1. How we protect data

  • Isolation per organisation — one customer’s data is not mixed with another’s
  • Authorized access — roles the customer sets limit who can see what
  • Audit — significant admin actions are logged
  • Encrypted transport — connections to users use HTTPS/TLS
  • Transcript only — we keep a text transcript of the conversation, not an audio/video recording
  • No external tracking on the website — no third-party analytics or ad pixels
  • MFA — two-factor sign-in (TOTP) for users

2. AI and human decisions

  • AI runs the interview and prepares a summary / evaluation for HR
  • There is no automated hiring decision and no reject email driven by a score
  • HR reads the report and decides

More: AI and regulation.

3. Hosting and sub-processors

Hosting is in the EU or in the customer’s region, depending on the agreement and availability. We use trusted providers to run the service (hosting, language/speech models, email, payments).

List: Sub-processors · processing on the customer’s behalf: DPA · privacy: Privacy policy.

4. Reporting a vulnerability

If you have found a security flaw, write to security@canvox.ai. Please give us a reasonable period to fix it before publishing, and do not access anybody else’s data while investigating.

Machine-readable contact details: /.well-known/security.txt (RFC 9116).

We do not pay bounties, but we answer every report and credit the reporter when they want to be named.

5. Contact

Security questions: info@canvox.ai