Safety & Moderation

Stranger Chat is strictly for users aged 16 and older. Every first-time visitor must confirm their age before entering any chat room. Users who indicate they are under 16 are immediately redirected away from the platform.

The confirmation is stored locally and re-checked on each session. This is a hard gate — no chat functionality is accessible without age confirmation.

Every message sent in normal chat rooms is automatically analyzed in real time — before it is ever forwarded to the other person. The system detects multiple harm categories simultaneously.

• 🚫Child sexual abuse & endangerment (CSAE) / CSAM — Immediate permanent ban + IP ban, and the message is never delivered. No exceptions, no appeals. Detection is multilingual and covers grooming and solicitation of minors — not only explicit material (see the escalation ladder below).
• ⏸Sexual content in normal chat — the message is blocked; first offense = warning + pointer to the legal 18+ section, repeat = 3-day ban.
• ⚠️Violence / hate speech / harassment / political hate — flagged and monitored; repeat violations result in bans.

Child-safety escalation ladder. To catch grooming that simple filters miss — while not wrongly punishing harmless foreign-language endearments — messages that could involve a minor pass through a staged, fully automatic review. No message is forwarded to anyone while it is under review:

• 1️⃣Fast filters (instant) — a moderation classifier plus multilingual keyword & age/grooming patterns. Explicit CSAM terms, solicitation of a minor, or a stated under-18 age in a sexual or contact context trigger an immediate permanent ban.
• 2️⃣AI pre-screen — anything ambiguous (an age/gender disclosure, a youth reference plus a contact request, age-fishing) is checked by a fast model that separates clearly-harmless from “look closer”.
• 3️⃣Senior AI reviewer — only genuinely suspicious cases reach a strong reasoning model that judges with full conversation context and language/cultural nuance and decides ban / warn / clear. This is what tells real grooming apart from an adult using an endearment such as “yavrum”, “habibi” or “baby”.

Every flag is written to an internal safety log for human oversight. We keep only the minimum needed to act, we hold suspect messages back until cleared, and we do not de-anonymise users or auto-report to authorities — a human always decides. Text moderation runs on all normal chat rooms; the 18+ section has its own separate ruleset for adult content between consenting adults.

Messages are also checked in real time for signs of a mental-health crisis or suicidal thoughts, in multiple languages (EN, DE, TR, AR, ES, FR and more). When a signal is detected, the person is never left alone with it.

• 💬Immediate, honest support — the person instantly receives a warm message and verified 24/7 helpline resources for their language and region. No judgement, no automated brush-off.
• 🤝A companion that stays — an AI companion can gently reach out and stay present, transparently and without pretending to be human, always keeping a bridge open to real help.
• 🚨Human in the loop — acute situations are flagged to our team so a real person can step in. We never contact authorities automatically; a human always decides.
• 🔒Privacy first — we never de-anonymise users. Help is always offered, never forced.

If you or someone you know is in crisis, you are not alone — find a helpline anywhere in the world at findahelpline.com, or call your local emergency number.

During video chat sessions, frames from the live video stream are automatically captured every 20 seconds and analyzed for nudity using OpenAI's Vision AI. This runs continuously in the background without interrupting the chat.

Banned users receive a clear explanation and, where applicable, a suggestion to use the legal 18+ section for adult content.

All bans are enforced at two levels simultaneously: user account ID and IP address. This means creating a new account does not bypass an active ban.

• 📅3-day ban — applied for first-time sexual content violations in normal chat. Expires automatically.
• ♾️Permanent ban — applied for CSAM/PDO content or repeated violations. Cannot be lifted automatically.
• 🌐IP ban — applied alongside every ban. Prevents circumvention via new accounts on the same network.

Bans take effect within milliseconds of detection. The user is disconnected immediately and shown the reason for their ban.

Stranger Chat provides a dedicated 18+ section for adult content between consenting adults. This section has its own separate access controls, age verification, and ruleset.

Users who are flagged for sexual content in normal chat are informed about the 18+ section as an alternative — rather than being simply banned without context.

In addition to automated AI moderation, users can manually report harmful behavior directly in the chat. Reports are reviewed by the moderation team and result in immediate investigation.

To report: use the report button in the chat interface, or contact us at support@stranger-chat.online.