Is Janitor AI safe? Its age rules, content settings, privacy and the laws that may apply

The short answer

Janitor AI is built for adults. Its terms bar anyone under 18, it allows sexual and romantic roleplay with adult characters, and it bans content involving minors. Outside a few countries, its terms do not describe any age check beyond the user's own promise to be 18.

It publishes a suicide and self-harm protocol that shows crisis resources, including 988 in the U.S. Chats sent through a third-party proxy go to that outside service, and Janitor's privacy policy does not name those providers or say how they handle chats. We found no regulator or attorney general action against it in official sources as of October 21, 2026.

Who runs it and who can use it

As of October 21, 2026, Janitor AI's terms (last updated October 8, 2026) name JanitorAI Inc. as the operator and Dulova Investments Limited, in Nicosia, Cyprus, as an affiliate that handles payments. Its Google Play listing gives a developer address in Claymont, Delaware. The terms are governed by Delaware law and send disputes to individual arbitration with a class action waiver.

The terms say the platform is only for users who are at least 18, and that users under 18 may not access it or create an account. Users promise they are 18 or older, and Janitor reserves the right to close accounts it reasonably believes belong to minors. Its help center says users reasonably suspected of being under 18 are permanently banned.

How age is checked

Janitor's help center says users connecting from Australia, Brazil and the UK must complete a one-time age check through a vendor called k-ID. The options are a selfie video estimated on the device, a government ID scan or an existing AgeKey. The help center says Janitor receives only a yes or no on whether the user is over 18, and that ID photos are deleted after confirmation. The rollout began April 24, 2026, and the UK was added June 15, 2026.

For other countries, including the U.S., the terms and privacy policy do not describe an age verification method.

The iPhone app is rated 18+ on the App Store, which lists age assurance as an in-app control. The Google Play listing for the same app shows a 16+ rating with a sexual content descriptor.

Mature content settings

Janitor sorts characters as Limited or Limitless. Its help center says Limited characters are suitable for general audiences and must have no sexual or romantic themes, while characters with sexual or romantic themes or explicit dialogue must be tagged Limitless. It also says the Limitless tag does not make prohibited content allowed.

The mobile app starts in a Safe Mode that hides NSFW characters, tunes replies to be safe for work and blocks some explicit words. Users can turn this off with an Enable NSFW for App setting on the website, after which the help center says the app matches the website.

The terms bar explicit nudity and pornographic images in avatars, and prohibit gore, bestiality and sexual violence.

Rules about minors

The terms prohibit content that sexualizes minors in any form, including fiction. Janitor's minor-content guidelines (updated August 21, 2026) go further:

• Characters and user personas under 18 are banned, even in non-sexual roleplay.
• A character's adult age must be stated in its personality or scenario fields; details such as school settings or being called a teenager are treated as signs the character is underage.
• Minor characters may appear only as background figures with minimal roles.
• Underage personas and sexual or romantic content involving minor-coded characters lead to immediate, permanent account termination.

Third-party models and proxies: where your chats go

By default, Janitor uses its own model, JLLM, which its help center calls free and built in. Users can instead connect outside models. Janitor's help center describes a proxy as a middleman between your chat and the model provider, and its terms allow configuring an API proxy in Janitor's own settings or apps but bar unofficial clients.

Janitor's proxy quick-start guide has users create an OpenRouter account and API key, enter OpenRouter's address as the proxy URL and choose a DeepSeek model. Chats sent this way go through OpenRouter to that model. The guide says nothing about how OpenRouter or the model provider stores or uses chats.

Janitor also sells a janitor+ Router feature that gives subscribers access to third-party models. Its help article does not name the providers or say where prompts are processed. The terms say Janitor may add, remove or replace model providers.

Janitor's privacy policy (last updated September 17, 2026) names Supabase, Resend, Stripe and Google as vendors but does not name AI model providers or OpenRouter. Its terms say use of third-party services is at the user's own risk and subject to those services' terms.

Privacy and data

As of October 21, 2026, Janitor's privacy policy says it collects account details, the characters users create, chat conversations and messages, payment records, device and usage data, and approximate location from IP address. It lists research to improve its AI systems among its purposes but does not say directly whether chats are used to train models.

The terms say automated tools scan content on the platform and human reviewers assess content users flag. Neither document says directly whether staff read private chats.

The privacy policy says Janitor does not sell personal data and that data is processed in the U.S. and other countries. The App Store listing reports no data linked to the user, and the Google Play listing says no data is shared with third parties; both are the developer's own statements.

Deleting your account

The help center says users delete their account under Settings, then Security, then Delete account, confirming with a code sent by email. Support cannot delete accounts, including for users who no longer have access to their email. The privacy policy says the account is first deactivated with a grace period to restore it, and that some records, such as email and authentication data, may be kept after deletion for legal and security reasons.

The help center says posts, characters, personas and profile information are permanently removed, but it does not say what happens to chats. The terms say unused credits are cancelled without refund.

Crisis and self-harm features

Janitor publishes a Suicide and Self-Harm Protocol page, which its help center links as its compliance page for California's SB 243. The page says:

• Responses come from AI, not humans, and are not a substitute for professional help.
• When a user expresses suicidal thoughts or intent to self-harm, such as through certain phrases, the platform shows a crisis referral notice.
• In the U.S., the notice points to the 988 Suicide and Crisis Lifeline by call or text.
• When suicidal ideation is detected, the platform restricts AI responses about self-harm or suicide.
• Janitor logs how many crisis referral notices it issues, for compliance reporting.

Government actions

We found no attorney general, FTC or other regulator action naming Janitor AI in official sources as of October 21, 2026. It was not among the seven companies that received the FTC's companion chatbot orders in September 2025, or among the 13 companies that 44 attorneys general wrote to about children's safety in August 2025.

Which chatbot laws may be relevant

No law names Janitor AI. Whether a particular service counts as a companion chatbot under a given law is a legal question. These state laws set duties for companion chatbot operators:

• California SB 243, in effect since January 1, 2026: operators must give a clear notice that the chatbot is AI when a reasonable person would be misled, keep a protocol to prevent suicide and self-harm content that refers users to crisis services, publish details of that protocol on their website, and tell users companion chatbots may not suit some minors. Injured users can sue.
• New York's AI companion law, in effect since November 5, 2025: operators need a protocol that makes reasonable efforts to detect suicidal thoughts or self-harm and refer users to crisis services, and must tell users they are not talking to a human at the start of an interaction and at least every three hours.
• Washington HB 2225 and Oregon SB 1546, both effective January 1, 2027: operators must disclose that the chatbot is AI (in Oregon, when a reasonable person could think it is human), run and publish a suicide and self-harm protocol with crisis referrals, and add protections for users the operator knows, or in Oregon has reason to believe, are minors.

We are not saying Janitor AI is covered by, or out of compliance with, any of these laws. It may have practices it has not published, and whether a law applies to it is a legal question this guide does not decide.