Taz

Safety on Taz

Last updated: 29 September 2026

Taz is designed with safety in mind. What you type is automatically checked for harmful topics, and common personal details are removed before it reaches the AI. The AI's answer is checked again before you see it.

No automated system is perfect. These checks reduce the chance of seeing something unsuitable, but they cannot rule it out, and AI answers can sometimes be wrong. If anything slips through, write to [email protected] so we can improve.

Automatic checks on each conversation

WhenWhat Taz is designed to do
Before the AI sees a messageCheck it for blocked topics and gently redirect the conversation
Before the AI sees a messageRemove common personal details such as email addresses, phone numbers, home addresses and card numbers
Before you see an answerCheck the AI's answer for unsuitable content
Before you see an answerRemove personal details that the first check may have missed

Topics Taz steers away from

Taz is set up to avoid violence and weapons, self-harm, requests for personal information or to meet in person, hate speech, extremism, and adult content such as sex, drugs, alcohol and gambling. When someone raises something worrying about their own feelings, Taz encourages them to talk to someone they trust. Taz is not a counselling or emergency service.

Breaks and screen time

By default Taz suggests a break after 60 minutes, with a reminder 5 minutes before. You can set your own daily time limit in Settings.

Your account

What happens to your data

Read the full Privacy Policy, or write to [email protected] about any safety concern. We aim to answer safety concerns first.

Open Taz

Labs to start with