Skip to content
Security and data

Data controls

The security settings admins control in TraceX, hiding personal data from AI models, fix approvals, share links, data retention and revoking access.

Admins find these under Security. Each one is a decision your security or compliance team may want a say in, so this page explains what each does and when to use it.

What it does: emails, phone numbers, card numbers, national ids, IBANs and IP addresses are replaced with placeholders such as «EMAIL_3f9a1c» before anything is sent to a model, and put back in what your team sees. Names, addresses and birth dates are hidden in named fields, senders and greetings. The same value always gets the same placeholder, so the model can still tell that two log lines are about the same customer.

Always on, whatever this setting: secrets, keys and tokens are always hidden.

When to turn it on: if your data includes customer personal data and your policies say it shouldn’t leave your systems unmasked. Answers stay readable because your team sees the real values.

What it does: members can write fixes and ask for a pull request; an admin approves it by opening it, in the dashboard or from Slack. High-risk fixes always need an admin, with this on or off.

When to turn it on: if you want a lead to see every automated change before it becomes a pull request. See Fixes and pull requests.

What it does: turns link sharing off for the organization. Every existing link stops working and nobody can make new ones. See Share a report.

When to turn it off: if reports must never leave the dashboard.

What it does: sets how long TraceX keeps investigations, traces and AI usage records. Data older than this is deleted in the daily cleanup. You can export everything first.

How to choose: match your company’s retention policy. Shorter retention means less history for Reports and for TraceX to learn from.

Settings → Revoke access disconnects TraceX from your systems at once. Use it in an emergency or during a security review.