Safety, stated plainly

What verification actually proves

Anyone offering to help sends us a government-issued document once, at the account level. A person on our team looks at it, confirms it belongs to them and shows an adult, then deletes the file — we keep the outcome, never the document. That check confirms they are a real, accountable adult. It does not confirm that the experience they claim is true — no identity check can do that.

So we don't say "our advisers are verified." We say: we verify who someone is, and we rely on monitoring and your reports to catch anyone acting in bad faith.

Monitoring is not optional here

• Every message in every conversation is screened by AI for crisis language and bad-actor patterns.
• Not sampled, not limited to people already flagged.
• Findings go into a review queue, not to a human reading your conversation for interest.
• You are told this is happening, in the conversation itself. No hidden surveillance.

Reporting is one tap

The report action is always visible inside a conversation. There's no multi-step form and no burden of proof on you. Every report is reviewed by a person, and every review ends in a recorded decision — a warning, a suspension, an escalation, or a dismissal as a false positive.

Four ways out, deliberately different

• Not a fit — before you've engaged. Silent, per-match, nobody is told anything.
• End the conversation, no fault — mid-conversation. Also silent. Most mismatches are nobody's fault.
• Block — account level, holds across everything in future.
• Report — goes to a human for review.

Keeping these separate is intentional. If every exit were treated as an accusation, nobody would use the gentle ones.

Crisis and clinical risk

Alongside deliberately does not match on clinical mental health conditions or active crisis. Every message is screened before anything else happens, and if the screen fires we stop and show real, region-specific resources plus encouragement to reach clinical support and people close to you. Your account is never blocked or penalised for it. High-severity signals alert a human immediately, not on a daily sweep.

What we don't collect

No medical records, no insurance documents, no diagnostic uploads. Other members never see your name, email or phone number — only the handle you chose.

What we haven't solved yet

• Someone blocked can, today, create a fresh account. We don't link identities across accounts yet.
• Age is self-declared at sign-up rather than document-verified.
• If someone is suspended mid-conversation, the other person currently isn't told — the thread just goes quiet.

These are known trade-offs at this stage, listed here rather than left for you to discover.