Talkie AI is one of the largest AI companion platforms in the world — roughly 11 million monthly active users, more than 734,000 Google Play ratings, and around 50,000 new characters created every day across 95+ languages. It is also the one most likely to already be on a child’s device without a parent knowing, for a reason that has nothing to do with its content.
So is Talkie AI safe? Two things make it a different problem from Character.AI or Janitor AI: what it collects, and how easily it routes around being blocked. We found data on both.
Is Talkie AI safe? The short answer
Not for children, and only conditionally for a supervised teenager. The content risk is real but roughly comparable to others in the category. What sets Talkie apart is the volume of data it collects, and the fact that the standard way parents block apps does not work on it.
The finding: a measurable bypass economy
Talkie was removed from the US Apple App Store in 2024. No public explanation was given. It returned in February 2025 under the name “Talkie Lab,” and the platform became fully accessible through any web browser.
That removal should have been a safety win. Our keyword data suggests it created a loophole instead. Using Google search-volume data via DataForSEO, we found the following monthly search demand:
| Search term | Monthly searches | Year-on-year |
|---|---|---|
| talkie ai online free | 2,900 | +175% |
| talkie ai now.gg | 2,400 | — |
| talkie ai unblocked | 880 | — |
That is over 6,000 searches a month for ways to reach Talkie without installing anything — and the middle row is the one parents should read twice.
now.gg is an app-streaming service that runs Android apps inside a web browser. The app never installs on the device. It executes on a remote server and streams to a browser tab. That means Screen Time app limits, Family Link app blocking, app-store age ratings and “check what’s installed on the phone” all see nothing at all — because nothing is there.
Meanwhile “talkie ai unblocked” is the term people use when something has already blocked it. School networks, mostly. That search exists because filters are working and are being routed around.
The honest caveat: search volume measures interest rather than confirmed users, and not everyone searching these terms is a minor. But the pattern is unambiguous — demand for browser-based, install-free access is substantial and the free-web-access term grew 175% year on year.
The other finding: what Talkie collects
When the Mozilla Foundation audited Talkie for its Privacy Not Included guide, it counted 223 data trackers firing within one minute of opening the app. That is far beyond what a typical social app loads in the same window.
A 2026 academic analysis went further, identifying 20 distinct advertising tracker SDKs embedded in the app — including Adjust, AppLovin and TikTok’s Pangle — while the privacy policy named none of them.
Talkie logs IP address, device model, operating system, mobile carrier, approximate location and in-app behaviour. It requests camera, microphone and storage permissions, in some cases when those features are not in active use.
Context matters for where that data goes. Talkie is published by SUBSUP PTE. LTD., registered in Singapore, and is backed by MiniMax, a Shanghai-based AI company valued at around $2.5 billion following an Alibaba-led $600 million funding round in March 2024. That ties the platform’s data handling to Chinese data law — a jurisdictional question that is separate from, and additional to, the ordinary privacy concerns.
For a teenager conducting emotionally intimate conversations, the combination of that telemetry volume and that data pathway is the strongest argument against the platform — stronger, arguably, than the content itself.
Teenager Mode and what it misses
Talkie offers a “Teenager Mode” intended to restrict mature content. It helps, and it is better than nothing, but it has documented gaps.
It does not fully block suggestive content. Because characters are user-generated, dialogue can drift into territory the filter does not catch — safety testers at Gabb Wireless found an innocuously themed character producing sexual innuendo almost immediately, without the platform having established the user’s age at any point.
Teenager Mode does remove user-to-user chat and following, which meaningfully reduces predation risk. But as Bark’s review notes, it does not remove the in-app link to Discord — which routes a supervised teenager straight onto an unsupervised platform. If Teenager Mode is enabled, that link is the gap to close, and our Discord parental controls guide covers the settings that matter there.
There is no supervised mode for underage accounts and no parent-facing dashboard. Conversations are not visible to parents in any form — there is no equivalent of Character.AI’s Parental Insights or Discord’s Family Center.
The stated minimum age has been reported inconsistently across sources, and in any case it is self-declared. As with Janitor AI, that is a liability disclaimer rather than a functional gate.
Regulatory attention
Talkie has not escaped official notice. The Australian eSafety Commissioner published guidance on the platform in September 2025, advising parents to review character types before permitting teen access. The platform was also referenced in the context of the September 2025 US Senate Judiciary Committee hearing on AI chatbot safety.
It is worth understanding why this has not translated into the kind of structural change Character.AI made. California’s SB 243 and New York’s AI Companion Models law impose real duties on companion-chatbot operators, but as of mid-2026 neither defines an actual age-verification standard — both require “reasonable” measures to identify minors. Operators facing wrongful-death litigation read that aggressively. Operators facing none have not.
How to actually block it
Given the now.gg route, app-level blocking is insufficient on its own. This needs to happen at the network and browser layer.
- DNS filtering at the router. A filtering DNS service applied at the home router blocks the domains for every device on the network simultaneously — including devices with no parental controls of their own. This is the layer that catches browser-based access.
- Block the app-streaming services too. This is the step almost everyone misses. Blocking the Talkie domain alone leaves now.gg and similar app-streaming sites open, and those are a working route back in. Block the streaming services as a category.
- Device-level content restrictions. Screen Time Content Restrictions on iOS and Family Link’s Chrome settings on Android restrict adult sites and support specific-site blocking. This layer follows the child off the home network.
- Both layers, always. Router filtering does nothing on mobile data. Device filtering does nothing on a device you have not configured. Families that do one find the gap quickly.
- Check installed apps anyway, since the Android app still exists and installs normally.
Where broader visibility across a device is warranted, parental monitoring software covers app and web activity generally rather than one platform — deployed openly, with the child’s knowledge. The bypass data on this page is the argument for that transparency: a teenager who is being blocked without a conversation searches for “unblocked,” and roughly 880 of them do so every month.
Common questions
Why did Talkie AI disappear from the App Store?
It was removed from the US Apple App Store in 2024 with no official explanation given, and returned in February 2025 as “Talkie Lab.” Access has been inconsistent since, with reported outages including a period in January 2026 — which is why “what happened to talkie ai” spikes periodically.
Does Teenager Mode make Talkie AI safe?
It reduces risk without removing it. Suggestive content still reaches users, and the Discord link remains active inside Teenager Mode, which is the most significant gap.
Can parents see Talkie AI conversations?
No. There is no parental dashboard, no linked-account feature and no supervised mode. Nothing in the platform surfaces conversations to a parent.
Is Talkie AI safer than Character.AI?
On age enforcement, no — Character.AI removed under-18 open-ended chat entirely in November 2025, while Talkie relies on self-declared age plus an optional Teenager Mode. See our reviews of Character.AI and Janitor AI.
The bottom line
Talkie AI’s distinguishing risks are data and reachability. It loads a volume of tracking on open that dwarfs a typical social app, it routes that data through a corporate structure spanning Singapore and Shanghai, and it named none of its advertising SDKs in its own privacy policy.
And because it now lives in the browser — reachable through app-streaming services that leave nothing installed — the parental control most families rely on cannot see it. A blocklist that stops at the app store stops short. For how the whole category compares, see our guide to AI companion apps.
