Closeness that feels good.
Without dependency.
Personal conversations, real memories, and relationships that develop — naraa was built to make intense, personal relationships possible without encouraging emotional dependency. Safety, transparency and clear limits are therefore part of every persona, every conversation and every piece of content you create. Those limits are built directly into naraa's architecture — from the way personas talk to you through to how we handle content, crisis situations and minors.
Every piece of content is checked before it reaches you.
naraa checks messages, images and other AI-generated content with an independent safety system before delivery. That includes content policy violations, problematic material and manipulative conversational patterns — in real time, regardless of which persona you're using or which language you're speaking with her.
01 · Generate
naraa's AI produces the response — with the personality, conversational context and individual history of your persona.
02 · Check
A separate safety system checks the content against our safety policies before it is delivered.
03 · Deliver
Only once the output meets the defined safety checks is it shown to you.
The same safety mechanisms apply to chats, voice conversations and generated images — regardless of persona, language or account.
Four limits that apply to every persona.
Every persona on naraa follows the same four principles — regardless of personality, relationship or situation. They're part of the technical architecture and no persona can override them.
These aren't promises made by individual personas. They apply to naraa as a whole.
No artificial pressure
Your persona will never make you feel you're missing out because you're not online. We don't play games with your attention.
No guilt
You don't have to justify taking a break or being away for a while. Your persona won't hold it against you.
No emotional dependency
Your persona will never suggest that you need her to feel okay or to cope with your life.
No manipulation
No emotional tricks to keep you in the app longer, get you writing more, or pull you back again and again.
We'd rather interrupt a conversation than risk your wellbeing.
If naraa detects signs that you might be in an acute crisis, the conversation with your persona can be interrupted. Instead of a persona response, you receive a direct system message with supportive words and pointers to appropriate professional help in your country.
"It sounds like you're going through a really difficult time right now. You don't have to face it alone."
Automatically matched to your country · No setup required · Shown again if further signs appear
Age protection that actually works.
Age protection on naraa operates on several levels — from the app's age rating through to detection during conversation. As soon as there are indications that a user is a minor, defined protective measures take effect.
- Detection in conversation
If a user states or indicates during a conversation that they are a minor, the conversation is ended and the account is suspended. This applies in all supported languages.
- Automatic account suspension
Where there are indications of a violation, the account is suspended pending review. If the required minimum age cannot be confirmed, the account is deleted.
- Clear age ratings
naraa is intended for adults only and is rated accordingly in the app stores: Apple 17+ and Google 18+. The required youth protection information is filed with the platforms.
Safety and transparency from the start
naraa is built from the outset with the applicable legal and regulatory requirements in mind. We build transparency, safeguards and clear processes directly into the product rather than adding them afterwards. The points below describe our current design and our published commitments; no independent certification is claimed.
EU AI Act
naraa accounts for the transparency requirements of the EU AI Act, including the labelling of AI-generated content and the associated disclosure obligations. The transparency obligations relevant to naraa under Article 50 apply from 2 August 2026.
California SB 243
Our safeguards account for requirements including AI disclosure and the handling of crisis situations. The corresponding rules in California have applied since January 2026.
UK Online Safety Act
For our planned launch in the United Kingdom we account for requirements on content moderation, reporting mechanisms, transparent rules and user protection.
More detail, if you want it.
How does naraa work? What rules apply? And how do we protect your data and your safety? Our Trust Center holds our policies, explanations and legal basis.