P
Pluck

Safety

Effective: 2026-09-04 · How Pluck handles crisis moments, content limits, and age rules.

Pluck is a practice space: you rehearse conversations with AI characters and get coaching from an AI coach. Practice can touch real feelings, so safety is built into every message — this page explains, in plain language, what we watch for and exactly what happens when it triggers.

Everyone you talk to is AI

Every character and coach in Pluck is artificial intelligence — there is never a real person on the other side. We tell you this before you can even sign in (the disclosure screen names our AI provider, Anthropic, and requires your explicit acknowledgment, which we record), and conversations are labeled as AI while you practice.

Adults only

Pluck is for adults 18+ only. Companion-style AI may not be suitable for some minors. The App is not directed to children, and the age gate is enforced on our servers: an account with an under-18 date of birth is permanently blocked, not just shown a warning it can tap past.

If a conversation shows signs of crisis

Every message you send — in a practice conversation, the texting simulator, real-world prep, or coach chat — is checked for signs of crisis before the AI responds. Two independent layers do the checking: a fixed, always-on check for unambiguous self-harm language that keeps working even if our AI systems are down, and an AI safety classifier for less direct signs of self-harm, suicidal thoughts, or acute distress.

If either layer detects a possible crisis, here is exactly what happens:

No automated system is perfect — a check like this can occasionally trigger on a message that wasn't a crisis, or miss one that was. If you're struggling, please don't wait for an app to notice: reach out to the services above, or your local emergency number if you're in immediate danger.

Content limits, in both directions

Sexually explicit and abusive content is blocked both ways. Messages you send are classified before the AI replies, and in role-play sessions repeated attempts to push disallowed content end the session. Replies from your conversation partner — in practice conversations and the texting simulator — and replies from the coach in coach chat each pass their own moderation check before they are shown or spoken; a generated line that fails is never surfaced. Other AI-written content in the App (coaching feedback, drills, prep plans) is produced by tightly constrained prompts rather than a separate output check, and every AI surface can be flagged from inside the App.

Reporting

Any AI reply can be flagged from inside the App. We review reports and use them to improve these systems.

Pluck is not a crisis service

Pluck is not therapy, counseling, or medical care, and it is not a substitute for any of them — it is a practice tool. If you are in immediate danger, call your local emergency number.

Questions

Safety questions or concerns: Studio Korvi inc. · contact@studiokorvi.com


Version française