Janitor AI Explained: What It Actually Is, What Its Own Rules Say, and Where UK Law Sits
What Janitor AI is in its own words, what its terms and privacy policy actually say, and how the UK Online Safety Act age-check duty applies to sites like it.
Janitor AI is a platform for creating and chatting with AI characters, aimed only at adults — its home page advertises “15M+ users” and its terms of service state it “is intended solely for users who are at least 18 years of age”. It is not a productivity assistant, it is not made by OpenAI or any other lab you have heard of, and its own legal documents never name the AI model that runs it. In the UK, the part that matters is the Online Safety Act’s age-check duty, and that is where this page spends most of its time.
Last checked: 22 September 2026.
Everything below was read from janitorai.com’s own published pages and from Ofcom and GOV.UK on 22 September 2026. One important limitation up front: this session’s browser and server are both outside the United Kingdom, so we could not test what a visitor from a UK connection actually sees. Nothing on this page claims to describe the UK user experience — only what the company publishes and what UK law requires.
What it is, in the company’s own words
The logged-out home page carries the line “Be anyone. Build anything.” and the claim “Join 15M+ users crafting living stories”. Its safety page opens by calling it “the world’s leading platform for interactive storytelling” and describes it as “a platform where millions of people create and interact with AI characters”.
The shape of the thing, from its own guidelines: users create characters, or “bots”, and other users chat with them. The guidelines apply “to all content, including characters/bots, images, profiles, and comments”.
The safety page sets out four commitments — privacy, creative freedom and community standards, trust and safety operations, and intellectual property — and this sentence sits under the second of them:
“Janitor is a platform for adults, and we respect their autonomy to tell the stories they want. We balance that freedom with a duty of care: zero tolerance for content that violates our guidelines, and a firm line against anyone suspected to be under 18.”
The three things people most often think it is
“It’s an OpenAI product” / “it’s a ChatGPT clone”. It is neither. It is an independent company — its guidelines and policies are published by JanitorAI Inc. — and it is not a general-purpose assistant. It is a character chat and creation platform.
“It’s a work tool.” No. Nothing in its own material describes a business, productivity or professional use. If you have landed here while looking for AI tools your business can actually run on, our guide to what AI tools actually cost a UK small business prices the mainstream, supported options from the vendors’ own pages, and the best AI tools for UK small businesses matches one to each business job.
“You can see what model it uses.” You cannot, from its published documents. We searched the terms of service, the community guidelines and the privacy policy for the names of AI providers, for “API key” and for “proxy”. None of them appear. The company publishes a great deal about moderation and very little about what is generating the text.
The UK position: what the law actually requires
This is the part that is missing from every other explainer, and it is the part a UK reader needs.
Ofcom’s guidance, published 16 January 2025, states the duty plainly:
“Robust age checks are a cornerstone of the Online Safety Act. It requires services which allow pornography or certain other types of harmful content to introduce ‘age assurance’ to ensure that children are not normally able to encounter it. Age assurance methods – which include age verification, age estimation or a combination of both – must be ‘highly effective’ at correctly determining whether a particular user is a child.”
And then, in a sentence written for exactly this category of service:
“Services that publish their own pornographic content (defined as ‘Part 5 Services’) including certain Generative AI tools, must begin taking steps immediately to introduce robust age checks, in line with our published guidance. Services that allow user-generated pornographic content – which fall under ‘Part 3’ services – must have fully implemented age checks by July.”
The headline on the same Ofcom page: “Pornography services must introduce age checks by July 2025 at latest”, alongside “Enforcement programme opening to monitor industry compliance”.
| What it covers | Deadline Ofcom set | |
|---|---|---|
| Part 5 service | Publishes its own pornographic content, “including certain Generative AI tools” | Begin taking steps immediately from 16 January 2025 |
| Part 3 service | Allows user-generated pornographic content | Fully implemented by July 2025 |
| All Part 3 user-to-user and search services | Must carry out a children’s access assessment | Final deadline 16 April 2025 |
A year and a half on, Ofcom reported to Parliament on how that is going. Its report on the use of age assurance, published on 16 July 2026 under section 157(5) of the Online Safety Act 2023, found:
- “Between July and December 2025, over 69 million age checks were completed across a sample of 32 services operating in the UK and analysed in this report – a 23-fold increase on the previous six months.”
- “All of the UK’s top 10 and the majority of the top 100 porn sites now have age checks in place.”
- But: “almost half of the pornography services visited by children have no age checks in place, and children can find these sites via search engines.”
- And a commercial consequence worth noting: “Many porn sites that have introduced age checks have experienced sharp declines in user traffic, while some sites without age assurance have gained in popularity.”
That report does not mention chatbots, AI companions or character-chat services anywhere. The regime plainly reaches them — Ofcom’s own guidance names generative AI tools — but the first six months of published evidence is about conventional sites.
What Janitor AI publishes about which countries it age-verifies in
Here is the finding, stated precisely, because precision matters more than drama.
Janitor AI’s safety page has a section headed “Regional information”, introduced with: “Some regions have specific legal requirements. Below is guidance for users and regulators in those jurisdictions, including how we verify age and comply with local online safety law.”
On 22 September 2026, that section named two jurisdictions:
| Jurisdiction listed | What the page says |
|---|---|
| Australia | “Compliance with the Online Safety Act 2021, mandatory age verification via k-ID, and reporting to the eSafety Commissioner.” |
| Brazil | “Compliance with the Digital ECA (Lei nº 15.211/2025) and mandatory age verification via k-ID.” |
The United Kingdom is not one of them. The words “Ofcom” and “Online Safety Act 2023” do not appear anywhere on that page. “United Kingdom” appears exactly once, in the mental-health helpline list at the foot of the page.
Its own FAQ is consistent with that. Asked “Is Janitor only for adults?”, the published answer reads:
“Yes. Janitor is a platform for adults. Any user suspected to be under 18 may be suspended or banned, and users in Brazil and Australia must complete age verification before accessing the platform.”
What that does and does not mean. It means the company’s published material describes mandatory age verification in two named countries, and the UK is not among them. It does not establish that a UK visitor faces no age check — a company is not obliged to describe every jurisdiction on a public page, and we could not test the UK experience from this location. Anyone who wants a definitive answer on that has the regulator’s route: Ofcom runs the enforcement programme and takes reports about services in scope.
What a UK reader should take from it is narrower and still useful: do not assume an age check is in place because the site says it is 18+. The site’s stated enforcement mechanism outside Australia and Brazil is suspension or banning of users “suspected” of being under 18 — which is a moderation response after the fact, not the “highly effective age assurance” that Ofcom’s guidance describes as the requirement before access.
What happens to what you type
Two of the company’s own documents should be read together here, because either one alone gives a misleading impression.
The safety page says: “We will never sell your personal data or share private conversations with third parties for advertising or commercial purposes. The conversations you have on Janitor are yours.”
The privacy policy — last updated 17 September 2026, five days before we read it — is more detailed. It states that what the company collects includes “User Content: Characters you create, chat conversations, messages, and other content you generate or upload.”
Its disclosure section permits sharing in three circumstances:
| Route | What the policy says |
|---|---|
| Service providers | “trusted third-party companies and individuals to perform services on our behalf, including hosting, analytics, payment processing, customer support, and marketing” |
| Business transfers | “In connection with any merger, acquisition, reorganization, sale of assets, or bankruptcy, your information may be transferred to a successor entity” |
| Legal requirements | Disclosure where it believes in good faith this is necessary to comply with “applicable laws, regulations, legal processes, or governmental requests” |
Those are not in conflict with the safety page’s promise — that promise is about selling data and about sharing conversations for advertising or commercial purposes — but they are a fuller picture, and the fuller picture is the one to make a decision on.
Retention. The privacy policy publishes no fixed period. It keeps data “only for as long as necessary”, judged against the duration of your account and active use, legal obligations, the need to resolve disputes or enforce agreements, and “legitimate business needs, such as maintaining security and preventing fraud”. In practice that means there is no date after which you can be sure a conversation is gone.
Your rights if you are in the UK. The policy’s European privacy rights section names the United Kingdom among the jurisdictions where users have GDPR rights, and states that you “have the right to lodge a complaint with your local supervisory authority” if you believe the processing of your personal data violates applicable law. In the UK that authority is the Information Commissioner’s Office.
What its own rules ban
The community guidelines, last updated 4 February 2026, are organised as seven blocks of “We do not allow:”. The first, and the one the company puts above the others, is about minors:
“Protecting minors is our highest priority. We do not allow any content that sexualizes, exploits, or endangers children. Users under the age of 18 are also prohibited from using our site.”
It requires characters to be “both physically and mentally over 18”, and prohibits romantic coding of underage characters as well as sexual content. The terms of service restate it in section 6, prohibiting content that “Depicts, promotes, or facilitates the sexualization of minors in any form, including fictional depictions” and any “depiction of individuals under 18 years of age”.
The terms also contain two things worth knowing before you sign up: an arbitration agreement and class action waiver in section 13, flagged in the terms’ own opening as something that affects your legal rights; and a disclaimer in section 11 stating the platform is provided “as is” and that the company does “not warrant the accuracy, completeness, or usefulness of any content on the Platform”.
The moderation figures the company publishes
These are the company’s own numbers, from its safety page, and we have no way to audit them. They are here because they are what it claims, and because a service’s published moderation scale is a fair thing to hold it to.
| Metric | Published figure |
|---|---|
| Items reviewed by AI | 100,000+ daily |
| Automated reviews | 1,000,000+ per week |
| Human moderation actions | 20,000+ per month |
| Tolerance for dangerous content | “0” |
The described process is two-stage: “Frontier AI tools trained on our Content Policy screen every piece of content before it is published”, then, if something is reported by a user, “Full-time moderators review the report around the clock and remove anything that breaks our rules.”
On industry work, the page states that Janitor AI is a sponsor of the Coalition for Secure AI (CoSAI), hosted by OASIS Open, and sits on its Project Governing Board.
If a chat has become a substitute for support
Janitor AI’s own safety page lists mental health helplines, including Samaritans on 116 123 for the United Kingdom. That is a reasonable thing for it to do, and it is worth extending.
If you are using a character chat because talking to people has become hard, the NHS route is direct and does not require a GP referral in most of England: our page on whether stress can cause physical symptoms sets out the NHS’s own wording on when to see a GP, when to call 111 and when to call 999, and how to refer yourself to NHS talking therapies.
What we left out, and why
- That Janitor AI is a “wrapper” routing to third-party models, or that users supply their own API keys. This is the most repeated technical claim about the platform. It may well be accurate. It is not stated in the company’s terms of service, community guidelines or privacy policy, none of which mention an API key, a proxy, a model provider or any named AI lab — so we are not asserting it.
- Prices. The terms confirm there are subscriptions, payments and a refund policy. We have not signed up, and we do not publish prices we have not read on a vendor’s own page on a stated date.
- What a UK visitor sees on arrival. Untestable from here, as set out at the top.
- Any judgement on whether the service is compliant with the Online Safety Act. That is a matter for Ofcom, which has an open enforcement programme. We report the law and the company’s published statements; we do not adjudicate between them.
- User counts as fact. “15M+ users” and “millions of daily active users” are self-reported and unaudited, and are described as such above.
- How to access the service from the UK, or anything resembling a workaround for an age check. Not something this page will help with.
If you came here looking for a tool for your business
This is a consumer service, and the licensing and data questions a business needs answered are different ones. If that is what you are actually after, our guide to free AI tools for UK small businesses sets out what each free tier allows, who owns the output under the vendors’ own terms, and whether your inputs train the model. Before any customer or staff detail goes into any AI tool, how to use AI in a UK small business sets out the checks the ICO expects.
Frequently asked questions
Is Janitor AI free?
Its terms of service include a section on subscriptions and payments and a refund policy, so at least some of the service is paid. We have not signed up or paid, and we do not publish prices we have not read on the company’s own page on a stated date, so there are no figures here.
Does Janitor AI have an age limit?
Yes, 18. Its terms state the platform “is intended solely for users who are at least 18 years of age” and that anyone under 18 is “strictly prohibited” from accessing it, creating an account or joining its communities. Its safety page adds that any user suspected to be under 18 may be suspended or banned.
Is Janitor AI safe to put personal information into?
Its privacy policy states that it collects your chat conversations and messages as User Content, and it publishes no fixed retention period — it keeps data by criteria such as account duration and legal obligations. It also allows for your information to be transferred to a successor entity in a merger, acquisition, sale of assets or bankruptcy. Those are reasons to treat anything you type there as something that may outlast your interest in it.
Who can I complain to in the UK about a site like this?
Ofcom is the regulator for the Online Safety Act, and it opened an enforcement programme to monitor industry compliance with the age-check duties. For data protection, Janitor AI’s own privacy policy states that users in the United Kingdom have the right to lodge a complaint with their local supervisory authority, which in the UK is the Information Commissioner’s Office.
Does Janitor AI sell my data?
Its safety page says: “We will never sell your personal data or share private conversations with third parties for advertising or commercial purposes.” Its privacy policy separately permits sharing with service providers who perform functions on its behalf, transfers in a business sale, and disclosure to comply with laws or governmental requests. Both statements are the company’s own; read them together rather than one alone.
How many people use it?
Its own home page says “Join 15M+ users crafting living stories”, and its safety page refers to “millions of daily active users across every time zone”. Those are self-reported figures published by the company, not independently audited numbers, and we have not been able to verify them.
Why do people say the characters forget things?
This is the most common complaint about the platform and we could not source a technical explanation from the company. Janitor AI’s terms, guidelines and privacy policy never name the AI model or models behind the service. Its terms do say the platform is provided “as is” and that it does not warrant “the accuracy, completeness, or usefulness of any content on the Platform”.
Sources
- janitorai.com — home page, safety page, community guidelines (last updated 4 February 2026), terms of use (last updated 29 July 2026) and privacy policy (last updated 17 September 2026), all read 22 September 2026
- Ofcom — Age checks to protect children online (published 16 January 2025)
- Ofcom — Report on the use of age assurance, laid before Parliament under section 157(5) of the Online Safety Act 2023 (published 16 July 2026)
- GOV.UK — Online Safety Act collection