AI Usage Policy

Even the best models occasionally invent facts, misread tone, or produce responses that are subtly wrong for a particular relationship. Always read a suggestion in full before you copy or send it.

You are responsible for every reply you send. Editing a suggestion, or choosing not to use one, is part of the workflow — not a failure.

• Legal advice, contracts, or official correspondence with authorities.
• Medical or mental-health decisions.
• Financial advice or trading decisions.
• Safety-critical situations (emergencies, threats, welfare or crisis support).
• Decisions where a wrong reply could cause harm to you or another person.

In these situations, speak to a qualified professional or the appropriate service.

Everyday messaging — WhatsApp threads, work chats, quick email replies, social DMs — where the goal is a natural, well-toned response in the right language. That is where we focus our quality bar.

We filter obviously harmful prompts and outputs (violence, sexual content involving minors, self-harm, malware, hate). Repeated attempts to bypass these limits may result in account suspension per our Acceptable Use Policy.

Suggestions are provided "as is" and are not guaranteed to be accurate, culturally appropriate, legally correct, or free from bias. Please verify before sending.