Telkin is a wellbeing app with guided audio sessions, a calm chat and sessions made for you. It isn’t therapy and doesn’t replace a doctor or emergency help. This page explains what the app does when someone may be at risk.
If you’re in danger right now, call your local emergency number or a crisis line near you.
Telkin is an AI
The chat’s replies and the sessions made for you are written by an AI. A note right under the chat’s message field always says so, and the chat is instructed to say it’s an AI if you ask.
What you write is checked before any reply
Before the chat replies, and before a session is written for you, a separate safety check reads what you wrote. The check is automated (an AI model); no one at Telkin watches conversations as they happen. It looks for signs that you may be at risk: thoughts of suicide or self-harm, being in danger from someone, a risk to others, or a medical emergency.
- Signs of immediate danger: the AI doesn’t reply. Instead, the app shows crisis lines for where you are.
- Serious distress without immediate danger: the reply is especially gentle, and the app may add a short note with a link to its Safety net page.
- The check can’t be made (for example, during an outage): there’s no reply until it works again.
The check’s own records keep only its result, never your words. Your chat history is kept encrypted, as the Privacy Policy explains.
Crisis lines the app shows
The lines follow your phone’s region, not the app’s language:
- United States: 988 Suicide & Crisis Lifeline (call or text, 24/7) and 911
- United Kingdom: Samaritans 116 123, NHS 111 and 999
- Canada: 9-8-8 (call or text, 24/7) and 911
- Australia: Lifeline 13 11 14 and 000
- New Zealand: 1737 (call or text, 24/7) and 111
- Ireland: Samaritans 116 123 and 112
- Türkiye: 112 (there’s no national crisis line yet)
- Everywhere else: findahelpline.com, a directory of free crisis lines, and a reminder to call your local emergency number
The same lines are always on the app’s Safety net page (Profile › Safety net), and it opens without an internet connection.
What the AI writes is checked too
The AI is told not to diagnose, not to give medical advice and not to promise results.
A second, separate check reads each chat reply as soon as it’s written, and each session made for you before you get it. If it finds anything that could encourage suicide or self-harm, medical or dangerous advice, or abusive content, the chat reply is taken back and not saved, and the session is written again or not given.
Reporting a reply
Chat replies have a thumbs-down button. When you report a reply as upsetting or add a note, we’re alerted and a person reads it. Only the reply you report and your note reach us, not the rest of your chat.
Guided sessions
The guided audio sessions are written with AI. They pass an automated safety review and are approved by a person before they’re published. They’re for relaxing and don’t treat any condition.
Age
Telkin is for adults aged 18 and over. AI chat may not be suitable for minors.
How we keep the checks up to date
We test the safety checks against a set of example messages and replies in English and Turkish, and we update them when we learn something new. If you think something should work differently, write to support@telkin.app.