Is Chai AI Safe? What the Chai App Shares About You

Chai AI is a legitimate product from a real, funded US company, and it is not malware. The safety question worth asking is about exposure rather than legitimacy, and there Chai's own filings say something no rival's do: in the data safety declaration Chai Research filed with Google Play, Chai states it shares your approximate location and the list of other apps installed on your phone with other companies for advertising or marketing. Its privacy notice is written to European law and never mentions California. And its own agreement asks you to confirm you have completed an age verification the product never asks you to do.

Everything below comes from Chai's own End User License Agreement, its own privacy notice, the Google Play data safety declaration Chai Research filed for its Android package, and Apple's live App Store listing, all read on 6 September 2026. Where a claim is reported by third parties rather than published by Chai, it says so.

Is Chai AI safe to use?

For an adult, yes, with one specific caveat about advertising data. Chai is published by Chai Research Corp., a Delaware registered corporation with file number 6527981, operating from 555 Hamilton Avenue, Suite 300, Palo Alto, California. That is a real, identifiable, US based counterparty, which is more than a large part of this category offers. Chai's corporate site states the consumer platform reached $70 million a year in revenue in 2026 on more than $55 million of total investment, and its privacy notice confirms data is processed and stored in the United States.

None of that is decoration. A funded US company with a published address is subject to US law, reachable by a US regulator, and far less likely to vanish overnight than the many companion apps that have shut down without warning. On the two questions people usually mean by "safe", is it a scam and will it steal my card, Chai comes out fine.

The caveat is what it does with data it collects legitimately, and that is a different question with a documented answer.

Is Chai AI safe for privacy?

This is where Chai separates itself from the category, and not favourably. Google requires every Android developer to file a data safety declaration listing what the app shares with other companies as distinct from what it merely collects. Chai Research filed the following for its Android package.

Read the second row again, because it is the unusual one. "Installed apps" is the inventory of other software on your phone. Shared with other companies. For advertising.

An installed app list is one of the more revealing things a phone holds. It shows which dating apps you use, which health apps, which banking apps, which recovery or mental health apps, which apps associated with a religion or a political affiliation. Handing that inventory to ad partners alongside an approximate location, from an account tied to your email address and your purchase history, builds a considerably sharper picture of a person than the chat logs everybody worries about.

For context, here is how the disclosure compares with the other companion apps whose filings we have checked.

Chai is the only one of the group declaring the installed apps share. Two points of fairness are owed here. First, Chai filed this itself, honestly, in a form Google audits, which is better behaviour than an app that stays out of the stores and files nothing. Second, this is standard mobile advertising plumbing rather than anything exotic; plenty of free apps outside this category do the same. It is simply worth knowing that the AI companion you talk to at midnight is part of that plumbing, and that its peers are not.

Chai's privacy notice is consistent with the filing rather than contradicting it. It states plainly that the company uses personal data to send advertisements to users on a free subscription, and that if you arrived through an advertisement it may share usage patterns with advertising partners for attribution. Chai also declares "Contains Advertising" on its App Store listing. The free tier is the ad supported version of the business, and Chai says so.

The practical defence is the same discipline that has moved into business software over the past two years, where teams stopped asking whether a vendor is trustworthy and started putting hard limits on what data an AI product can actually reach. On a phone, the consumer version is short: deny location permission to a chat app that has no use for it, since approximate location cannot be shared if it is never granted, and turn off personalized advertising in your Android or iOS settings so the identifier the share is keyed to stops being useful.

How long does Chai keep your chats after you delete your account?

Up to five years. Chai's privacy notice states that data is retained for no longer than necessary, "not exceeding, in any case, the duration of the account and five (5) years following its deletion", with a longer window possible where Chai needs it to defend legal claims or comply with an order.

Deleting the account is genuinely easy: sign in, open Profile, then Settings, then Delete Account. What that button does is end the account. It does not end the retention. Five years is not unusual as a legal ceiling and it is more transparent than the many apps that say nothing at all, but it is a long time for conversations people treat as ephemeral, and almost nobody reads far enough into a privacy notice to find it.

Does Chai AI verify your age?

No, and the wording of its own agreement is odd about it. Section 8 of Chai's End User License Agreement, updated 28 July 2026, states that Chai is "strictly restricted to users aged 18 and older" and then asks you to acknowledge that you are legally an adult and "have provided accurate age verification".

No age verification is collected. Signup takes a self reported age. Apple rates the listing 17+, which is a store level classification rather than a check on you, and Google Play applies its own rating. So the agreement has a user affirming a step the product never performs, and the gap between an 18+ contractual claim and a 17+ store rating is not resolved anywhere.

That is not unique to Chai, and it is worth putting in context rather than treating as a scandal. Almost no US facing companion app runs real age assurance, because no US federal law currently requires it, and the platforms that do check identity almost always added it for the United Kingdom, Australia or a specific set of US states rather than for the country as a whole. The pattern, and which apps sit on which side of it, is set out in our look at Character AI age verification and what it actually asks you for, which covers the same question on the largest platform in the category.

Is Chai AI safe for kids?

No, and Chai's own documents say so more clearly than most parental control articles do. The EULA restricts the platform to adults. Apple's rating is 17+. Chai's self declared content advisories to Apple are frequent and intense mature or suggestive themes, frequent and intense profanity or crude humour, and mild sexual content and nudity, alongside references to alcohol, tobacco and drugs, horror themes, cartoon violence and, on the Google Play listing, guns or other weapons. The listing also declares user generated content and advertising.

User generated content is the part that matters for a minor. Chai's library is built by its community, and the volume is very large, which means moderation is reactive by design. Chai's answer to that is an in app reporting tool and a public safety centre, and its acceptable use rules are strict on paper: Section 7 prohibits sexually explicit images and videos of any kind, child sexual exploitation material, terrorist and extremist material, and content encouraging self harm or suicide, with violations resulting in immediate account termination, data preservation and potential reporting to law enforcement.

Those rules are enforced after the fact, on a platform with no age gate. A determined 15 year old is not stopped by a checkbox. Parents evaluating this should treat the 18+ claim as a contractual term rather than a control.

Is Chai AI actually uncensored?

Less than its reputation suggests, which cuts both ways depending on what you want. Section 7.1 of the EULA states that the generation, sharing or distribution of sexually explicit images or videos of any kind is strictly prohibited on the platform. Chai's App Store advisories match that: frequent and intense on suggestive themes and profanity, but only infrequent and mild on sexual content and nudity. By comparison, PolyBuzz declares frequent and intense sexual content or nudity to the same store.

So on Apple's own scale, Chai is a more restrained product than several of its rivals, and buyers arriving expecting an unfiltered platform are working from reviews rather than from the rules. Whether that reads as safer or as a limitation is entirely down to what you are shopping for.

Is Chai AI legit, and who owns it?

Chai is legitimate. Chai Research Corp. is a named Delaware corporation, file number 6527981, at a real Palo Alto address, publishing a dated EULA (28 July 2026), a dated privacy notice (2 July 2026), enforced store disclosures on both platforms, and business figures on its corporate site. That is a stronger paper trail than most of this category manages.

One inconsistency deserves flagging for American readers specifically. Chai's privacy notice is drafted to the GDPR and the UK GDPR. Its data subject rights section opens by telling you the listed rights apply "if you are a citizen of the EU or the UK". The words California, CCPA and CPRA appear nowhere in the document. A California headquartered company has written its consumer privacy notice to European law and given the residents of its own state no stated rights section at all.

Californians hold rights under the CCPA and CPRA regardless of what a policy says, so nothing is lost legally. What is lost is the instruction manual: a UK user is told exactly how to exercise access, erasure and objection, and a Californian has to work it out. When a company's disclosure quality is this uneven, the honest reading is that compliance was bought for the market that fined people first.

What does Chai cost, and does paying change any of this?

Paying removes the advertising you see. It does not remove the data sharing you do not see, because the Google Play declaration covers the app rather than a tier.

Chai's pricing is also more complicated than any review admits. Its App Store in app purchase list shows six separately priced products: Premium Membership at $9.99, Premium Subscription at $13.99 a month or $134.99 a year, Ultra Subscription at $19.99 a week or $29.99 a month, and a $2.79 consumable pack. The weekly Ultra plan is the one to watch, because $19.99 a week is $1,039.48 across a year against $359.88 for the identical monthly plan. We broke down all six tiers, the arithmetic and what each one actually unlocks in our Chai AI review covering premium, subscription and pricing.

Chai's EULA, for its part, contains no clause about subscriptions, renewal, cancellation or refunds anywhere in its 21 sections. Its only mention of a refund directs you to the store operator. Cancel through Apple or Google, not through Chai.

The honest summary

Chai is safe in the sense that matters least and exposed in the sense that matters most. It is a real US company, it takes payment properly, it will not steal from you, and it publishes more than most of its rivals. It also tells Google, in writing, that it shares your approximate location and your installed app list with other companies for advertising, keeps your data for up to five years after you delete your account, writes its privacy rights section for Europeans, and asks you to certify an age verification that does not exist.

If the trade you are making is "free companion, ad supported, I understand I am the product", Chai is an honest version of that deal and its disclosures let you verify it, which is rare. If you would rather not be in that trade at all, the deciding factor is what an app shares rather than what it collects, and we lined the whole category up on exactly that question in the best AI companion app for privacy.

Or start with the demo at the top of this page. It is one companion who remembers you, built for adults, with no ad tier underneath it and no installed app list going anywhere.