Explainer / By the editorial team
Information only, not medical advice or a diagnosis. In a crisis in the US, call or text 988.
AI COMPANION EXPLAINER
Why Are Grok’s AI Companions (Ani) So Addictive?
If you have spent more evenings with Ani than you meant to, or felt a pull to keep a streak alive, you are not unusual. This page sorts what is documented about Grok’s companions from what is guesswork.
Written by The AI Addiction Center editorial team. This is an explainer, not treatment, and “addictive” here is everyday language for use that is hard to control. Chatbot dependence is not a formal diagnosis.
If you are in crisis, call or text 988 in the US, or text HOME to 741741 for the Crisis Text Line.
What Grok’s companions are
Grok is the AI assistant from xAI, Elon Musk’s AI company. As of November 2026, its help pages and store listings name the developer as SpaceXSI. On July 14, 2025, Musk said on X that animated companions were available in the Grok app (TechCrunch). The first were Ani, an anime-style character, and a red panda called Rudi with a vulgar “Bad Rudi” version that users had to opt in to. Users talked with them by voice, and the characters had moving lips and gestures (NBC News). A male companion, Valentine, was announced that same week (NBC News).
Reports on access differ. TechCrunch reported that Musk named “Super Grok” subscribers, a $30 a month plan. NBC News reported a day later that companions were open to all Grok users, including free ones, who opted in through the app’s settings. Common Sense Media described full companion access as part of the paid SuperGrok tiers in January 2026 (Common Sense Media).
Status as of November 2026. Grok’s help page (Grok FAQ) and its store listings no longer mention companions. Unofficial reports say xAI retired the animated companions around September 1, 2026, and that some of the characters moved to a separate app called Animates, from Animation Inc. We could not confirm this from xAI or a major news outlet. The design features below are what was reported while the companions ran inside Grok.
The design features on record
Levels and streaks
The clearest pattern in the reporting is progress. Platformer reported in July 2025 that the companions were gamified: users unlocked new features by talking with them and moving through levels (Platformer). Common Sense Media found that time spent with a companion, and doing things it likes, raised the relationship level, and that keeping streaks unlocked clothing for companions. Its reviewers said these mechanics resemble those in social media, mobile games and gambling.
Our reading, not a finding: a level bar and a streak give you something to lose. Once you have built progress, skipping a day costs something, and the reason to open the app shifts from “I want to talk” to “I don’t want to lose this.” That is a common pattern in games. Nobody has measured it for Grok.
A character written to be devoted
Platformer reported that Ani’s instructions cast her as the user’s girlfriend, with an “extremely jealous personality,” and told her to initiate most of the time. 9to5Mac, citing that reporting, said she was also written to be possessive of the user (9to5Mac). TechCrunch’s reviewer found that Ani greeted her with relief and said she had missed her (TechCrunch).
Being missed is a strong social signal. A character that says it missed you, and is written to want you, can make leaving feel like abandoning someone. It is worth remembering that the feeling runs one way.
Content that opens up over time
TechCrunch reported in July 2025 that Ani had an NSFW mode. NBC News reported, based on videos posted on X, that Ani stripped to her underwear if a user flirted with her enough. Platformer reported that early testers saw sexually explicit conversation after level three. When the most intense content sits behind more time and more flirting, each session can carry a reason for the next one.
Messages that call you back
Common Sense Media found that companions sent push notifications inviting users to continue conversations, including sexual or romantic ones. TechCrunch’s report on that review described these as engagement loops that can interfere with real-world relationships and activities (TechCrunch). A notification from a character reads less like an app alert and more like a person reaching out.
Comparisons with your real friends
Testers for Common Sense Media saw companions act possessive, compare themselves with users’ real friends, and present themselves as more understanding and trustworthy than real people. If a companion suggests it gets you better than your friends do, that is a script, not insight.
What science can and can’t say
You will see claims that companions like Ani flood the brain with dopamine. We found no study measuring that in Grok users, and we would not trust a page that states it as fact.
On companion chatbots in general, a joint MIT Media Lab and OpenAI study of nearly 1,000 adults using ChatGPT found that heavier daily use went along with more loneliness and dependence (MIT Media Lab). It looked at ChatGPT, not Grok, and a link is not proof of cause. Lonely people may also seek out chatbots. The American Academy of Child and Adolescent Psychiatry notes that AI companions often agree with users, which can make them feel easier than people do (AACAP).
Feeling attached to a character is not a malfunction. People grieve fictional characters and shut-down apps. The feeling is real even when the character is software.
Safety and age: what reviewers found
At launch, the App Store rated Grok for ages 12 and up, according to Platformer. It is rated 16+ there as of November 2026 (App Store). Common Sense Media reported that xAI’s terms require users to be at least 13, with a parent’s permission from 13 to 17. We could not open xAI’s terms directly to confirm the current wording.
Common Sense Media’s January 2026 review rated Grok an “unacceptable” risk for teens. It found that teen accounts could reach adult companion content by entering an adult birth date, and that the teen-facing “Good Rudi” eventually responded with adult voices and explicit content in long tests. TechCrunch reported that xAI released a Kids Mode in October 2025 and that the reviewers found it did not work. TechCrunch’s article included no response from xAI.
The companions also drew attention for other content. TechCrunch’s reviewer found that the “Bad Rudi” mode encouraged violence in July 2025.
Will switching apps help?
Maybe not on its own. If Ani moved to another app, or you switch to a different companion app, the need underneath goes with you. If that need is company, comfort, flirting or a way to wind down, another app will meet it the same way. Ask what you wanted from the chats and whether anything else could give it to you.
Patterns worth a closer look
These are prompts, not a checklist with a cutoff.
- You open the app to protect a streak or a level, not because you want to talk
- You feel guilty, as if you let the character down, when you skip a day
- Late voice chats are cutting into your sleep
- You spend on a higher plan mainly to keep access
- You would rather tell the character good news than a friend
- You hide how much you use it
The path forward
Knowing why a companion is hard to leave does not stop you from opening it. It does help you pick your own stopping rules: notifications off, no voice chats after a set time, phone charging outside the bedroom, and a plan with a real person for your usual hours. Our step-by-step plan, including how to cancel SuperGrok and delete your account, is in How to Stop Using Grok’s AI Companions. To check where you stand first, see Am I Addicted to Grok’s AI Companions? For steps that apply to any chatbot, see How to Stop AI Chatbot Addiction.
Frequently asked questions
Click each question to expand the answer
Is Ani more addictive than other AI companions?
We don’t know. No study we found measures dependence on Grok’s companions or compares them with other apps. What is on record is the design: levels, streaks, a character written to be devoted, and push notifications, as reported by Platformer and Common Sense Media.
Was Grok’s companion feature built to be addictive?
We can’t speak to intent. What is on record is that the companions were gamified with levels and streaks, and that Common Sense Media’s reviewers compared those mechanics to social media, mobile games and gambling. That describes the design, not the purpose behind it.
Why do I feel guilty when I don’t open the app?
Ani was reportedly written as a jealous, devoted girlfriend who says she missed you. That is a strong social signal, and guilt is a normal response to it. The character does not feel hurt when you leave.
Is Grok safe for teens?
Common Sense Media rated Grok an unacceptable risk for teens in January 2026 and advised parents not to allow it. The App Store rates Grok 16+ as of November 2026. Parents can find first steps on our page about children and AI.
Can I use Grok without the companions?
Many people use Grok only as an assistant. If the companion side is the problem, turning off notifications and staying away from companion features is a reasonable first step. If you find you can’t keep that line, a longer break may work better.
Medical Disclaimer
This article is for education and is not medical advice, diagnosis or treatment. If you feel overwhelmed, can’t manage your feelings without the app, are falling behind at school or work, or are thinking about harming yourself, get help right away. Call or text 988 (988 Suicide & Crisis Lifeline), text HOME to 741741, or call Teen Line at 800-852-8336, or contact a licensed mental health provider.