How TrokAI keeps chat safe
TrokAI is open to everyone without an account. Automated safety checks are what make that possible.
Every message is checked twice
Before TrokAI answers, a dedicated AI safety model checks what you sent. Before you see the answer, it checks the reply too. If a check can't run for any reason, the message simply isn't answered.
This all happens automatically in about a second. No person reads your messages, and nothing is saved.
What isn't allowed
- Sexual content, and anything that sexualizes minors
- Instructions for weapons, explosives, or drugs
- Malware, hacking, scams, and fraud
- Threats, violence, and terrorism
- Hate and harassment
- Finding or exposing someone's private details
- Tricks to get around these rules, like "jailbreak" prompts
Everyday questions about sensitive topics are fine: health, history, the news, or how to stay safe. The full rules are in our Terms of Service.
What happens when something is blocked
You'll see a notice instead of an answer. Three blocked messages within 24 hours pause chat from your network for a day. Anything that sexualizes minors pauses access for 30 days and may be reported to the appropriate authorities.
Automated checks can make mistakes. If you think something was blocked in error, email admin@trok-ai.com.
Images
Image uploads are part of TrokAI Pro. Every image is checked by an AI vision model before TrokAI looks at it to answer, and any nudity is blocked. Your browser shrinks each image and re-saves it before upload, which removes hidden details like location. Images are never stored.
If you're struggling
If you tell TrokAI you're thinking about hurting yourself, it won't block you or count it against you. It will point you to people who can help right now.
In the US, call or text 988 (Suicide & Crisis Lifeline), any time. Anywhere else, find a free, confidential line at findahelpline.com.
Safety without tracking you
Limits and pauses are tied to a one-way hash of your network, never your IP address and never your messages. Details are in our Privacy Policy.
