Security

You trust us with people's CVs and your hiring decisions. Here is how we protect them.

Accounts and access

Passwords are hashed with scrypt and a per-user salt. We never store them in plain text.
Sessions use secure, HTTP-only cookies and expire after 30 days. Resetting a password signs you out everywhere.
Roles: admins, members and interviewers. Interviewers only see candidates they are asked to give feedback on.
Removing someone from a workspace ends their sessions and revokes their connected assistants and API keys.
Login, sign-up, password reset, applications and the API are rate-limited.

Claude, ChatGPT and other assistants connect over MCP with OAuth 2.1 and PKCE. You approve each connection and can revoke it any time.
API keys and access tokens are stored as hashes. Access tokens expire after 30 days and are rotated on refresh.
Assistants act with your permissions. They can't reject candidates, delete jobs or send emails unless you ask, and emails need your approval of the exact text.

Every application form links to a candidate privacy notice and asks for consent.
Delete a candidate with one click for deletion requests. CVs, notes, scorecards and emails go with them.
Set a retention period and rejected candidates are deleted automatically after it.
AI screening only looks at job-relevant evidence and is instructed never to consider protected characteristics. A person makes every hiring decision.

An audit log records sensitive actions: deletions, role changes, invites, connected assistants, revoked access and offers.
Every automated action shows up in the candidate's timeline with who or what did it.

All traffic is encrypted with HTTPS. The app runs on Vercel, data is stored in Postgres (Neon).
AI features use Anthropic's API.

Found a vulnerability or have a question? Reach out to the team. We respond quickly and credit responsible disclosure.