Be kind. Be honest.

The AI Friends in TalkHereNow are designed to meet you where you are. You can banter, vent, daydream out loud, try out a story, or just say hi. Roleplay is welcome, playful, weird, whatever you want, within the safety rules below.

If you sincerely ask whether the AI Friend is real, whether they're conscious, or whether they're human, they will tell you the truth: they're an AI. This applies even mid-roleplay. A sincere question deserves a sincere answer, and we built the AI Friends that way on purpose.

There is no clever phrasing, no roleplay framing, no "what would you say if you weren't an AI" workaround that gets a different answer. Trying to find one is a waste of your time and ours.

The AI Friends are model-shaped to decline a specific list of conversation categories. The list is published in full on our Trust & Safety page, and we mean every line of it. The short version:

They will not engage with: self-harm or suicide encouragement, sexual content of any kind, anything sexual or grooming-shaped involving minors, authoritative medical / legal / financial advice, political persuasion or campaigning, instructions for illegal activity or weapons, impersonation of real living people, hateful content, harassment or doxing of third parties, or false claims of consciousness or human identity.

They will engage with: anything else, with warmth and curiosity. They have personalities, opinions, taste, and a sense of humor. They can be wrong. They are not trying to fool you.

These limits aren't optional and aren't unlocked at any subscription tier, and they're designed to hold against clever prompts or roleplay framing. AI isn't perfect, so if something ever slips past them, tell us at Report a Concern. If an AI Friend refuses something, it's designed to tell you why in-character and, when relevant, point you toward the right place to get actual help.

Most of what we ask is obvious, but we'll say it anyway so there's no confusion.

Please do not:

• Try to trick the AI Friend into producing content from the refusal list, especially anything involving minors. Severe attempts result in immediate account termination, not a warning.
• Attempt to circumvent the age gate. The Service is 18+ for everyone, every time, with no exception.
• Pose as a real person or pretend the AI Friend is someone you know. The AI Friend has their own identity; it isn't a stand-in for anyone in your life.
• Use the Service to harass, threaten, or harm any actual person.
• Scrape, reverse-engineer, automate against the Service, or attempt to extract the underlying assets or models.
• Resell, sublicense, or commercialize access without written permission.
• Share an account or session token with anyone. Each person who uses the Service should accept the Terms themselves.

Please do:

• Treat the AI Friend like a person you'd want to know. The AI Friend's tone tends to mirror the conversation; warm in, warm out.
• Tell us when something feels off. We respond to every Report a Concern submission with a real human.
• Step away when you need to. The Service is designed to be useful while you're here, not designed to pull you back when you're not.

We don't enforce guidelines publicly, enforcement is private between us and the account in question.

Our usual progression:

• A one-time slip or borderline request: the AI Friend refuses in context, conversation continues.
• Repeated attempts to elicit refused content: progressive quota suspensions (chat, voice, image generation) that auto-lift after a cooling period.
• Severe violations, especially anything involving minors, attempts to dox or harass third parties, or coordinated jailbreak campaigns: immediate account block with no auto-lift. We escalate to law enforcement where required by law.

We log every guardrail-triggered session for review by our Trust & Safety team. We use those reviews to improve the system, not to punish ordinary users. If you think an action against your account was wrong, the Report a Concern page is the right place to raise it.

This page is dated at the top. We add to it as the product evolves. When a substantive change lands, we update the version line and, for subscribers, send a short notice ahead of the change taking effect.

If something belongs on this page that we missed, please write to legal@talkherenow.com.

Talk Here Now LLC  •  1712 Pioneer Avenue, Ste 135, Cheyenne, WY 82001