Is Janitor AI Safe? Privacy, Age Checks and Proxies (2026)

Adults only. Janitor AI is an 18+ platform. Its Terms of Service say users must be at least 18. This page is written for adult readers and contains no explicit content.

How this page was made. This is a documentation-based safety review built from Janitor AI’s safety page, Privacy Policy, Terms of Service, Suicide and Self-Harm Protocol, newsroom and changelog, all read on 18 October 2026. AITried has not audited Janitor’s systems, and a policy describes commitments rather than proving practice.

The short answer

Janitor AI’s written commitments on privacy are fairly strong for a free companion platform: no selling of data, no sharing of private conversations for advertising, and self-service deletion. Its main risks come from how people use it. Many users connect third-party models through proxies, which sends chats to companies Janitor does not control, and the 18+ rule is self-declared in most countries.

Who runs Janitor AI

The Privacy Policy, last updated 17 September 2026, names JanitorAI Inc. It says the company is based in the United States and processes data there and in other countries. Payments may be processed by an affiliate, Dulova Investments Limited in Nicosia, Cyprus, and by third-party processors such as Stripe.

The official app launched on app stores in February 2026. Days earlier, Janitor warned on 4 February 2026 about fake apps on app stores. Use only janitorai.com or the app linked from it; our guide to spotting fake AI app clone sites explains how to check.

What data Janitor AI collects

The Privacy Policy lists:

• Account details: email, username, password and profile information.
• User Content: characters you create, chat conversations and messages.
• Payment information, processed by payment providers.
• Usage data, including pages visited, time spent, click patterns and search queries.
• Device information, IP address and approximate location from your IP.
• Cookies and similar tracking for sessions, analytics and recommendations.
• If you sign in with Google: your Google account ID, email, name and profile picture.

How it is used

Listed uses include operating and improving the platform, personalising recommendations, moderation, and “conducting research and development to improve our AI systems and user experience”. The policy does not spell out whether chat logs specifically are used to train Janitor’s own model (JLLM), and we did not find a training opt-out. We could not verify either point.

Sharing and selling

The safety page states: “We never sell your personal data or share private conversations with third parties for advertising or commercial purposes.” The Privacy Policy lists sharing with service providers (hosting, analytics, payments, support and marketing), for legal requirements, with consent, and in a merger, acquisition or bankruptcy, where data may pass to a successor.

The proxy question: the biggest practical risk

Janitor lets users chat with its own model or connect outside models. The help centre includes guides to proxy setup and error guides for third-party services, and the July 2026 Router launch post acknowledges that “many people already connect external models to JanitorAI through proxies”.

When you use a third-party model, your messages go to that provider. That provider’s privacy policy, logging rules and training terms then apply to your conversation, not only Janitor’s. Janitor’s Privacy Policy does not describe how proxy providers handle chat content.

The paid janitor+ Router, launched 30 July 2026, removes the need to paste API keys into external proxies, which reduces one risk (handling your own keys). It still sends requests to third-party model providers, and the Terms say Router charges are based on “usage reported by the underlying model provider”. For what Router costs, see is janitor+ worth it.

Before connecting any external model, check:

1. Does the provider log prompts, and for how long?
2. Does it train on API traffic?
3. Where is it based, and can you delete your data?

Age rules and verification

The Terms say anyone under 18 is “strictly prohibited” from using the platform, and the company may terminate accounts it reasonably believes belong to minors.

Verification depends on where you connect from:

Janitor’s announcement says k-ID offers on-device facial age estimation, an ID scan deleted after the check, or a reusable AgeKey, and that Janitor receives only a yes or no on whether you are over 18. These are the company’s statements. The same post says the system “relies entirely on the IP address you connect with”, which means it can be bypassed. For households with teenagers, that makes device-level controls important.

Moderation and reporting

The safety page says:

• Automated systems trained on the Content Policy screen every piece of content before publication.
• Full-time human moderators review reported content around the clock.
• It cites over 20,000 human moderation actions per month and over 1 million automated reviews per week. These are Janitor’s own figures, not independently verified.
• Content involving minors, real individuals, or the glorification of self-harm is not allowed.

To report something, tap Report on any character or comment, open a ticket at help.janitorai.com, or email support@janitorai.com. Rights holders can use the DMCA process. Janitor also publishes a Complaints Policy and a Content Removal Policy.

Crisis protocol and where to get help

Janitor’s Suicide and Self-Harm Protocol states that responses are AI-generated, “are not human”, and “are not a substitute for professional help”. When a user expresses suicidal ideation or self-harm intent, it says it shows crisis resources and restricts related AI responses.

The safety page lists:

• United States: call or text 988 (Suicide and Crisis Lifeline).
• United Kingdom: Samaritans on 116 123.
• Australia: Lifeline on 13 11 14.
• Elsewhere: findahelpline.com.

If you are in immediate danger, call your local emergency number.

Deleting your account

The Privacy Policy says to go to Settings, Security, Delete Account. The account is deactivated first; a changelog entry describes a 30-day period during which you are signed out, your profile is hidden and your characters go private, and you can restore the account by signing in. After permanent deletion, some records (such as your email and purchase records) may be kept. The Terms add that deleting your account cancels any unused Router credits without refund.

Practical steps if you use Janitor AI

1. Use a nickname and a separate email address.
2. Keep names, addresses, workplaces and health details out of chats.
3. Read the policy of any proxy or model provider before connecting it.
4. Never paste API keys into unofficial sites or apps.
5. Delete your account, not only the app, when you stop.

Our AI companion privacy checklist gives twelve questions to ask any platform. Compare with is Candy AI safe and is Character.AI safe.

What we could not verify

• Whether chat logs are used to train Janitor’s own models, and whether an opt-out exists.
• How proxy and Router model providers store or use chat content.
• The privacy contact address, which was obscured on the page we read.
• Moderation figures beyond Janitor’s own statements.

Bottom line

Janitor AI makes clear, adult-focused commitments: no selling of data, no sharing of private chats for advertising, round-the-clock review of reports, and self-service deletion. The weak spots are the self-declared age rule in most countries and the third-party models many users connect, which take chats outside Janitor’s policy. Adults who stick to the built-in model or vet their provider, and who keep identifying details out of chats, reduce most of that risk.