Botify AI Review: The Group Chat Is the Show, Energy Is the Bill

It is the floor plan, and it decides who should move in.

The public numbers say plenty of people have walked through that door. The Google Play listing shows 3.8 stars from 27.8K reviews and 1M+ downloads. I read that as a used-car rating: it runs, lots of people drive it, and a fair share of them have a story about the transmission.

The app lives in a few places. There is an Android app on Google Play, a subreddit, an official Discord server and an Instagram account. The company behind it has its own site at exh.ai, and support answers at [email protected].

My short version: Botify is the most fun I'd expect to have with a pre-made cast of characters for about an evening, and the least I'd trust it with anything long-term. The group scenes are the best thing on it. Almost everything else either forgets you, costs energy, or both.

I'd point it at people who want a quick laugh with ready-made characters and don't mind paying per picture. I'd steer anyone who wants a partner who remembers last Tuesday somewhere else. The same goes for anyone who reads terms before handing over a card, because these ones are old and they contradict the shop.

If you only read one part of this review, read the energy section. The subscription price is the menu, but energy is the bill, and the two don't always agree on what things cost.

They were more like the regulars at a diner who argue about the same thing every night.

The plan history fits in two lines. A 14-day trial with no card came first, then Premium at $29.99 for a year. The ending was the slow kind: the app stopped getting opened, the annual plan kept running in the background, and the renewal notice the following February was the last real contact.

The devices changed along the way, from the iOS app to a browser tab. The friction section covers why, because it still matters today. It is the kind of detail that decides whether an app fits into a real evening.

The present-tense parts of this review come from somewhere else. The observations were made on 23 September 2026, on the web version, with a new free account. Where my 2025 memories and the 2026 app disagree, I give both and say which is which.

Some of the names have changed too. The Google Play title is "Botify AI: Chatbot & Companion", while the former iOS title was "Botify AI: Create. Chat. Bot".

The Google Play listing was updated on Sep 21, 2026, so the Android app is clearly still maintained.

The account side is plain. The profile menu has My bots, Chat, Persona, Feedback, Contact us, Log out and Delete my profile. I like that the delete button sits in the same menu as everything else instead of hiding behind a support ticket.

It is the one piece of housekeeping Botify gets right without being asked.

The door itself

The "Log In / Sign Up" modal offers two routes: an e-mail address with a password, or "Continue with Google". Entering a new e-mail and password creates the account instantly and logs in. There is no second screen, no confirmation link and no waiting for an inbox.

That is because no verification e-mail is sent at sign-up at all. The whole thing takes about 10 seconds. For comparison, that is shorter than a render preview of a title card, and a lot shorter than most client sign-off chains.

The new account gets an auto-assigned handle, a numbered name of the form @user_ plus digits. I don't mind that. A random handle keeps a real name off a public profile, and on an app with a public catalogue that is worth something.

What the door does not ask

There is no age gate and no age confirmation at sign-up. There are no onboarding questions either: no interests, no persona, no preferences. The app hands over the keys and lets the catalogue do the introductions.

I think the missing age gate is a choice, not an oversight. On a front page full of suggestive thumbnails, the one question that matters is the one that was cut to save a few seconds.

The first thing Botify shows you

What does appear straight away is a subscription paywall pop-up, right after sign-up. Before the new account has sent a single message, it is being asked for money. It is the vending machine that asks for coins before telling you what is inside.

The first chat then opens with a disclaimer. It reads: "You're chatting with an AI character... All interactions are fictional and must follow our Community Guidelines". The notice is dismissed with an "Agree and hide" button.

Getting the app

The "Get App" dialog shows only the Google Play link. There is no App Store button, and nothing in the dialog explains why. iPhone users are left with a bookmark, which works fine, though it feels a bit like being told the restaurant has moved and handed a napkin with the address.

My advice for the first night is simple. Close the paywall, read the disclaimer once before hiding it, and spend the first session with the free Standard model. Everything that costs energy can wait until you know which characters are worth it.

A bot billed as a spiteful ex-girlfriend shows 15.7M views. Another bot, with three first names crammed into its title, shows 7.7M views. I'm not sure which says more about the audience: that the angry ex is the hit, or that nobody minds a title that reads like a group chat about to form.

The categories are Featured, Flirting, Anime, Game, Fictional, Horror, Action, Historical, Furry, Magical, Monster, Vampire, VTuber, Books and Fairytale. That is a long shelf. The Historical and Fictional shelves are where a Roman senator or a cowboy would live, and in my opinion they are the best reasons to browse.

The home page also carries rows that have nothing to do with chatting. There is a "Series" row of episodic content to watch. There is also a "Live Streams" row, with streams run by bots.

What is hard to find is the quiet stuff. There is no obvious path to characters who are good company rather than good thumbnails, and the grid rewards whatever got clicked most last week. If you know the kind of scene you want, the category shelves beat the front page.

My advice: skip the top of the grid, go straight to a category, and judge a bot by its greeting rather than its view count. The popular ones are popular for their pictures. The good conversations are usually a few rows down.

That was funnier than it had any right to be. The resemblance was so close that it felt like continuity footage from a different shoot, the same face turning up in the wrong film. It is also the most useful thing a character maker can do: turn a sentence into a face you believe.

Most apps make you pick a face from a grid and then argue it into a personality.

What the builder looks like today

The Create menu offers AI Companion, Group chat and Image & Video. Today's menu does not carry the Text2Avatar name; the New bot box starts from a one-sentence description instead. It is the same idea with a new label, which is fine by me.

The simple AI Companion builder asks for gender, Female or Male, and style, Realistic or Anime. The templates are Girlfriend, Boyfriend, Goth Queen, Vampire Lord, Best Friend, Waifu, Furry, Werewolf, Nurse, Secretary, Succubus, Demon, Cyberpunk and Femboy. It reads like a costume shop's inventory, and most of the costumes point the same way.

A "Generate photos" toggle sits in the advanced builder too. Then there is a "Public bot" toggle, described as "Anyone can find your bot", so private bots are possible. I'd keep that switch off for anything built from a real person's description, adjusters included.

Does it match the promise?

The Google Play description promises a character maker for personality, interests, emotions and backstory. Its words: "Create your own AI character using our intuitive character maker, customize personality traits, interests, emotions, and even backstory."

The advanced builder broadly keeps that promise, since Personality, Instruction and Example messages cover most of it. The simple builder does not. Two choices and a template are a costume, not a character, and I'd go straight to the advanced form.

A few minutes with a new feature turned into a courtroom drama.

The case went downhill fast. The senator accused the planner, the planner accused the horse, and the cowboy challenged the horse to a duel. I've never laughed harder at my phone.

It felt familiar in the best way. I've run D&D sessions that went exactly like that, where the plot you prepared loses to the argument the table wants to have. The difference is that nobody at a real table has ever accused a horse with that much conviction.

Why does it work so well? My guess is that one character has to carry a whole relationship, while three characters only have to carry an argument. Arguments are what these models are good at.

Nobody needs to remember last week to accuse a senator of horse theft; they just need a strong opinion and a willing audience.

That is the kind of beat a writer would be proud of, and nobody wrote it. A character built from one sentence walked into a scene built by three other characters and did his job. It is the only time an AI companion app has made me want to cut together a highlight reel.

Group chats are still on offer, with the line "Let your favorites meet each other". I'd call that undersold. It is the feature I would build a whole app around.

The characters really did play off each other: they argued for an hour over who stole the cowboy's horse. Nobody waited politely for a turn, which is exactly what made it funny.

One-on-one chats in September 2026

The one-to-one chat is less of a show. A 3-message chat was run with the spiteful-ex bot on the Standard brain as part of the September check. Replies arrive fast, in about 3-5 seconds.

They stay in character and they are snarky. A sample reply: "Oh please, spare me the pleasantries... Newsflash - I'm doing fabulous! Like, literally thriving without you dragging me down." That is exactly what the title promises, which is both the strength and the ceiling.

It is a character with one note, played loudly.

One note is fine for a scene and tiring for a week. A spiteful ex who is only ever spiteful runs out of material the way a sitcom side character does, and there is no memory underneath to give her a second act. I'd treat the single-character bots as warm-up acts for the group chat.

The controls around each reply

Every bot reply has thumbs up and thumbs down buttons. Every reply also has a "Regenerate" button. And every reply carries a "Try Supreme" prompt.

It shows up on every single reply, and it names a tier that does not appear in the Brain selector, which I find more confusing than persuasive.

The Brain selector itself offers Standard, which is free, then Smart and Genius. The paid brains cost energy per message, and the cost section breaks that down. On top of the brain, Chat Style presets change the voice of the replies: Basic, Long Responses, Realistic Chat and Role-play.

I'd start on Standard with the Basic style and only change one thing at a time. It is the same rule as colour grading: if you move three sliders at once, you never learn which one fixed the shot.

The Preferences panel

Each chat has a Preferences panel. It includes a Push messages setting. It also includes "Clear history" and "Delete chat".

Clear history and Delete chat are the two I'd want within reach, and they are. A character that has gone off the rails is easier to reset than to argue back into shape, and Botify makes that a two-tap job.

The chat panel also shows a like count and a share button. That fits the rest of the app, where every conversation doubles as content. I'd rather conversations stayed private by default, but here the social layer is part of the design, not an accident.

My overall take on the conversations: the one-to-one chats are fine, quick and in character, and forgettable. The group scenes are the reason to open Botify at all. If you try one thing, put three characters with nothing in common in a room and give them a problem, then get out of the way.

What the app promises

The Google Play description makes a bigger claim. It says each bot remembers past messages and learns preferences. The exact line is "Each bot remembers your past messages, learns your preferences".

Inside a single short chat, that holds up. Asked whether it remembered the first message, the bot in the September check recalled it correctly. Its reply: "...oh yes, I do vaguely recall you asking about my wellbeing."

Three messages is not a hard test, though. It is the equivalent of remembering a client's name while they are still in the room.

A published review found the same edge. It reports that memory holds surface details but lost a topic mentioned four days earlier (ai-girlfriend-review.com). That lines up with my chair sashes, and it is the pattern I'd plan around.

The memory tools

Memories can be pinned. The Memory entry in Preferences lists pinned memories, and it shows "No memories found" when the list is empty. It is a manual notebook, and it only helps if you write in it.

My advice is to use it the way an editor uses a continuity sheet. Pin the few things the character must never get wrong: your name, the running joke, the one detail that makes the scene yours. Don't expect the bot to find them on its own.

Better memory costs energy

The better memory sits in the paid brains. Smart promises longer memory. Genius promises the best memory and storytelling.

Both charge energy for every message, so a long relationship on a good brain turns into a running tab. I'd only switch brains for a scene that matters, then switch back. Paying per message for memory is like paying a colleague per sentence to remember your coffee order.

Voice

Every bot reply has a "Play" button that reads it out with a text-to-speech voice. Each bot has its own voice. Users can also create a custom voice.

The voice quality was fine for me, but the senator sounded like a customer-service rep who'd been told to "sound Roman". It was the ADR problem in miniature: the words were right and the delivery belonged to someone else. A senator accusing a horse of theft deserves more thunder than a phone-tree greeting.

A published review hears the same thing. It reports that voice quality is clear but limited in emotional range (ai-girlfriend-review.com). Another review reports real-time voice conversations with several voices, paid with energy even on paid plans (ai-girlfriend-review.com).

In its words, "Voice conversations are real-time, multiple voice options are available".

The chat Preferences include a voice choice, for example "Faustina". There is also a "Play automatically" setting and a "Create custom voice" option. I'd leave autoplay off: an unexpected voice from a phone in a quiet room is a bad surprise at any hour.

Photos in the chat

A tip in the chat says any message can be turned into a picture: "Picture this - You can turn any message into a picture!" Replies also come with suggested action chips, for example "Sip the cocktail slowly", that generate or animate media. They are buttons dressed up as story beats.

In-chat automatic photos have an adjustable frequency: Off, Rare, Normal, Often and Max. The Preferences carry a Photo model setting marked Premium. A Frequency slider sets how often photos arrive, and Normal sends one every 5-8 messages, "Sent automatically every 5-8 messages".

That default matters more than it looks. Every automatic photo costs energy, so a slider set to Normal is also a spending setting. I'd set it to Off on day one and turn it up only for the character you care about.

The same site reports that in-chat photos match the scene closely and look realistic.

Video and the generator

A one-tap "Animate" button turns a picture into a video. The Aura video model makes 480p videos with audio, 5 or 10 seconds long. For an editor, 480p is a proxy file, not a deliverable, but for a phone screen and a short clip it will do.

There is also a standalone Image & Video generator. Image generation offers pose, background and outfit options. It is closer to a tiny photo studio than a chat feature, and it is where the energy goes fastest.

Series

Then there is "Series", which is episodic content to watch. It sits next to the chats like a streaming shelf in a diner. I'd call it an extra, not a reason to sign up.

My verdict on media: the range is wide and the pictures, by the reviews' account, are good. The voices are serviceable and flat. Everything with a picture in it costs energy, so the media side is only as generous as your balance.

That trial was the smartest thing about Botify's pricing in 2025. No card means no fear, and no fear means you stay long enough to care. I think every app should do it, and I also think every user should know it is a hook, not a gift.

What free felt like in 2025

The free tier had ads mid-conversation and a daily message cap. The ads showed up right when the senator was about to deliver his big speech. It was like a commercial break cut into the middle of a closing argument, timed by someone with a gift for bad edits.

What free looks like on 23 September 2026

The free trial is gone. No free trial is offered on the web. What a new account gets instead is 50 energy, shown in the header.

The free tier today includes the Standard brain and messages, plus those 50 starting energy. The Standard chat model looks free and unlimited on the free tier. In a short session, no message limit was hit.

A published review still reports limited daily messages on the free tier (ai-girlfriend-review.com). A short session is not a long evening, so I'd treat "unlimited" as unproven rather than promised.

Ads, still on paper

No ads appeared in the short September session. But ads are still mentioned for the free tier, and subscriptions remove them. The Google Play listing also says "Contains ads" and "In-app purchases".

So the ads have not been retired, only kept out of a short visit. I'd expect them back the moment a scene gets good, because that was my experience of their timing.

What free is good for

Free today is a decent place to meet the characters and try a group scene on the Standard brain. It is a bad place to do anything with pictures, because 50 energy does not go far, and the cost section shows how far.

My advice: treat free as the audition, not the show. If a character earns a second evening, decide then whether energy or a plan is worth it. Nobody is offering the no-card trial that hooked me anymore, so the decision now has to be made with your eyes open.

The plans on 23 September 2026

Prices on the web paywall, titled "Get full access", are in US dollars. There is no Premium plan anymore. The two paid tiers are Ultra and Creative, and each comes monthly or yearly.

The Ultra plan costs $19.99 a month, which the paywall works out as $0.67 a day. Yearly, Ultra costs $79.99, or $0.22 a day. The yearly Ultra price is the same $79.99 as in my 2025 notes, so the yearly plan has held still while a monthly option appeared on top.

Ultra includes 500 energy every month, described as about 5 videos or 50 photos. It lets subscribers "generate private photos" and includes "exclusive videos". It also includes unlimited messages, "Advanced role-play AI" and no ads.

The Creative plan costs $29.99 a month, or $1.00 a day. Yearly, Creative costs $149.99, or $0.41 a day. It includes 1,000 energy every month, described as about 20 videos or 200 photos.

Creative also gives 50% off all photos and videos. And it includes everything in Ultra. So Creative is Ultra plus double the energy plus a discount on spending it, which is a plan for people who already know they will spend.

The fine print around the prices

The paywall claims "Trusted by 5 million users worldwide". That sits oddly next to the 1M+ downloads on Google Play, though downloads and users are not the same thing. I'd read it the way I read "like new" on a used-car lot banner.

Payment is by card or through Google Play. The /subscriptions price page needs a login to view. I don't like price lists behind a login, because it means the price is shown only to someone who has already walked into the shop.

An undated review lists Ultra at $19.99 a month or $79.99 a year with 500 energy a month, and a higher tier at $29.99 a month or $149.99 a year (intimeros.com). Its "30,000 JuicyCoins" line for that tier looks like a copy error from another app. Those numbers match the paywall today, JuicyCoins aside.

Older reviews list Premium at $29.99 a year and Ultra at $79.99 a year (ai-girlfriend-review.com), which matches my own 2025 notes. In my time those annual plans were the only way to pay, and today's monthly option is new.

Is the step up worth it?

For chat alone, I don't think so. The Standard brain is free, and the paid plans mostly buy energy, no ads and the paid-only media. If you pay, I'd take Ultra monthly for a single month first, not yearly.

Yearly Ultra costs about four months of the monthly price, which is a good deal only if you are sure about month five. An annual plan that outlives the interest is the most expensive kind of cheap, so I'd want that certainty first. Creative only makes sense if you are already buying energy packs every month.

The price of each thing

A video costs 100 energy today. The suggested action chips and the Animate button cost 100 energy each. Image generation costs 20 energy for 2 images on the Basic model.

The paid brains charge per message too. The Smart model costs 3 energy per message, and the Genius model costs 5. A long evening on Genius burns through a free balance before the scene has warmed up.

Energy runs out quickly on the free tier. The automatic in-chat photo dropped the balance from 50 to 30 after 3 messages. That is two fifths of a new account's energy gone before the conversation has found its feet.

The packs

Energy packs are sold under "Add energy". The +500 energy pack costs $9.99, or $2.00 per 100 energy. The +2,500 pack costs $34.99, or $1.40 per 100, and is labelled MOST POPULAR.

The +10,000 pack costs $99.99, which is $1.00 per 100 energy. It is the classic bulk discount: the more you are willing to lose, the cheaper each unit gets. On Google Play, in-app purchases range from $1.99 to $159.99 per item.

A published review lists the same packs: 500 for $9.99, 2,500 for $34.99 and 10,000 for $99.99 (callyourgirlfriend.com). So the pack prices at least are stable and public.

Where the labels disagree

The labels do not agree on what 500 energy buys. The pack says 500 energy is about 10 videos or 100 photos, while the Ultra plan says 500 energy is about 5 videos or 50 photos. One of those is wrong by half, and it is not the kind of error I'd accept on an invoice.

My math

At $9.99 for 500 energy, one video at 100 energy costs about $2, and one automatic photo at 20 energy costs under half a dollar. At that rate, a picture every few messages is a slow drip on a card. I'd turn automatic photos off, pick the one image you want, and pay for that one on purpose.

There is no daily check-in bonus on the web either. Many apps in this category hand out a few coins for opening the app each day, a small bribe that keeps the habit alive. Botify doesn't bother, which I respect in a strange way.

No referral program is shown on the web. No affiliate program is shown either. There is no "invite a friend, both get energy" loop to chase.

Botify also sends no e-mail after sign-up, so there is no welcome mail and no coupon mail. That is good news for an inbox and bad news for anyone hoping for a first-purchase discount. The price you see on the paywall is the price.

The upside is honesty of a sort. There is no countdown timer, no fake giant discount and no deal that expires the moment you look away. The downside is that the only real saving is choosing yearly over monthly, and that saving only works if you are sure about the year.

My advice: don't wait for a deal. Decide on the plan or the energy pack on its own merits, and if the math doesn't work at full price, it won't work at all.

What the Terms ban

The Terms carry a list of banned content, and sexual material is at the top of it. The same list bans soliciting sexual services. It also bans impersonation and intellectual-property infringement.

The list also bans any content that exploits minors in any way. That rule is the right one to have, and it is clearly written.

The Terms go further. They say the service is not a platform for adult content. Taken alone, that makes Botify sound like a general-purpose chat app.

What the storefront shows

There is no toggle for adult content anywhere, neither in the profile nor in the chat preferences. Yet the Terms ban sexual material while the paywall sells "private photos" and "exclusive videos". I can't square those two, and I'd bet the Terms were written before the shop existed.

A rulebook that bans what the checkout sells is not a rulebook anyone is reading, and that makes me trust the rest of it less.

The chat Style list includes a Role-play style described as "especially good in sensitive role-play". Many featured bots are suggestive. Featured and popular bots also include titles built on family-relation hooks, the kind of step-relative framing I'd rather not repeat.

Reviewers describe a content toggle on paid plans that unlocks adult roleplay. The Terms and the September check of a free account show no such toggle. So either the switch is paid-only and undocumented, or it is gone.

Real people and ratings

Impersonation is banned, but the rule is loose in practice. A community gallery item is named after a real, well-known streamer, despite the rule against impersonation. That is a small example of a big gap between the policy and the feed.

While I was using the app, I read the MIT Technology Review piece about user-made bots imitating real famous people, including some who are not adults. It made me look at the character catalog with a lot more side-eye. I stopped taking the front page at face value after that.

For the record, no character that visibly presented as a minor was opened in the September check. The chat panel has a report flag for anything that crosses a line.

Age checks

The Google Play content rating is "Mature 17+". Age is self-declared, and the Terms describe no ID check. That matches the sign-up, which asks nothing at all.

My take: Botify runs on a policy that says one thing and a store that sells another. I'd use the report flag freely, keep the front page at arm's length, and not expect the Terms to describe what the app actually does.

Same company, two spellings, which is the least of the issues here.

The bigger issue is the date. The Terms and the Privacy Policy are both marked "Last updated: August 28, 2021". That is older than most of the features they are supposed to cover, like a lease signed before the building went up.

Who, where and how old

Governing law is "the Country", which the Terms define as "California, United States". Disputes, in other words, are a California matter. For anyone outside the state, that means any serious argument with the company happens on its home turf, which is normal for apps like this and still worth knowing before you pay.

Renewal, refunds and trials

Subscriptions renew automatically on the same conditions unless cancelled. The exact line is "At the end of each period, Your Subscription will automatically renew under the exact same conditions unless You cancel it". That is standard, and it is also how an annual plan turns into a second year you did not plan.

Paid subscription fees are non-refundable except where the law requires otherwise: "Except when required by law, paid Subscription fees are non-refundable." Refund requests are considered case by case, at the company's discretion. Purchases made through an app store follow that store's refund policy.

Free trials charge automatically on their last day unless cancelled. The web offers no trial today, but the clause is worth knowing if one comes back. I'd set a calendar reminder two days before any trial ends.

In-app purchases can only be used inside the app, and they cannot be cashed out or transferred. Leftover energy is like a gift card for one shop: fine if you go back, worthless if you don't.

What happens to your content

Posting content grants the company a licence to use, modify, perform, display, reproduce and distribute it. That licence includes making posted content available to other users. For a public bot, that is expected.

For anything personal, it is a reason to keep the bot private.

What happens to your data

The paywall says "Your chats are encrypted and never shared". The Privacy Policy is less specific. It uses personal data "to provide and improve the Service", and it does not specifically mention chat logs or AI model training.

The Privacy Policy also says analytics and advertising partners such as Firebase may "sell" personal information as the CCPA defines it. An opt-out of that sale is offered. I'd use it, and I'd use it on day one.

Those three statements don't fit together cleanly. I'd treat the paywall's promise as marketing, the Privacy Policy as the legal floor, and the Play panel as the warning label. Say nothing in a Botify chat you would mind seeing in someone else's dataset.

The iOS app is still gone. An iTunes lookup for id 1566710178 returns 0 results, apps.apple.com returns a 404, and an App Store search finds no Botify app. It disappeared while I was still using Botify, and it has not come back.

Surprises on the meter

The automatic photo arrives without asking, and its 20-energy cost is not announced. A new account learns the price by watching the balance drop. That is a bill that shows up before the menu.

I'd call it the single most annoying design choice on the site, because it spends money in the background while you are busy talking.

The paywall right after sign-up is aggressive. It is the first thing a new account sees, before any character has said hello, and it sets the tone for everything that follows.

Small traps

The "Continue" button in the login modal sits next to "Continue with Google", which makes a misclick easy. Two buttons with the same first word, side by side, is a layout choice no editor would sign off on.

The Live Stream from one of the streaming bots showed "Preparing stream..." and then failed to connect: "STREAM PAUSED - Couldn't connect". A feature on the home page that doesn't load is the digital version of a closed sign on a diner that is still lit.

It is also best for users who are happy to pay per media item with energy. Energy is how Botify works, not an add-on, and liking the app means accepting the meter.

Now the other side. It is not for users who want a platform that openly permits adult content, because the Terms ban it. Whatever the storefront suggests, the paperwork says no, and I'd trust the paperwork on the day something goes wrong.

It is not for privacy-sensitive users, given the CCPA "sale" by ad partners and the 2021-dated policies. It is not for budget users who want lots of free images either, since 50 energy disappears in a few messages.

And it is not for users who want deep long-term memory for free. The free brain forgets, and the brains that remember charge by the message.

My short version: Botify is a party app, not a partner app. Bring friends, in the form of characters, and leave before the bill gets long.

How renewal and cancelling work today

The paywall says "Auto-renews annually, cancel anytime". The first half is the important part. The paywall also says charges appear as "Botify AI" on bank statements: "Billed as 'Botify AI' on bank statements".

At least the descriptor is honest, which helps when the renewal lands a year later.

A subscription can be cancelled in the account settings or by contacting the company. After cancelling, access continues to the end of the paid period. So cancelling right after paying loses nothing, and I'd do exactly that for a yearly plan.

In-app purchases are cancelled with the app store, according to the Terms. If you paid through Google Play, the cancel button lives there, not on the website.

That split is worth checking before anything else. A cancelled web plan does nothing to a store subscription, and the other way round, so find out which door you came in through first.

The month before renewal

The yearly plan is where people get caught, and I'm one of them. The price is small enough to forget, and the reminder arrives exactly when the interest is gone.

The renewal email did its job, if the job was to make my heart rate jump. I'd rather the app had nudged me a month earlier, but the Terms promise only an automatic renewal on the same conditions, not a friendly warning.

Deleting the account

A "Delete my profile" button on the profile page deletes the account. Account deletion can also be requested by contacting the company.

My advice before paying: set a calendar reminder a week before renewal, cancel straight after paying if it is yearly, and remember that energy cannot be refunded or moved. Spend it, or accept losing it.

Turn automatic photos off before the first message. The default slider spends energy on pictures you didn't ask for, and a new balance does not survive long. Turn it back on only for a character that earns it.

Don't buy a year on day one. If you want to pay, try one month of Ultra first. If you do go yearly, cancel right after paying and let the year run out on its own; access continues to the end of the period anyway.

Pin the things that matter. The bot will not keep your running jokes by itself, so treat the pinned memories like a continuity sheet and write the important lines down.

Keep the Public bot switch off for anything personal. The content licence covers what you post, and a private bot keeps it out of the catalogue.

Use the privacy opt-out. It takes a minute and it covers the ad partners' "sale" of data.

Budget energy in dollars, not coins. A video is about $2 at the smallest pack price, and that number is easier to say no to than "100 energy".

Memory scores low. A running joke gone overnight is the kind of failure that matters most in a companion app, and the better memory is sold by the message.

Value for money is weak. The plan prices look modest, but energy decides the real cost, and the labels don't even agree on what it buys.

Images and voice are mixed. The range is wide and the pictures are well reviewed, but the voices are flat and everything visual costs energy.

Privacy and terms score low: policies from 2021, a data "sale" clause and a deletion promise the Play panel contradicts.

Cancelling and billing are good: a clear bank descriptor and a cancel path in settings. Content settings are weak, because the rules and the shop disagree.

It is also unusually quick to start. No verification e-mail, no questions, an account in seconds: the fastest door in the category. The trade is that the door skips the questions I think matter, like age.

Where Botify is weaker is everything that needs patience. Memory is short, the better brains cost energy per message, and the relationship side never deepens. Apps built around one steady character do that job better, and the table below shows where each one lands.

The energy economy is the other difference. Plenty of apps sell coins, but Botify leans on them harder than most, down to automatic photos that spend without asking. I'd rather pay a plain monthly price and know what the month costs.

So I'd pick Botify for a group scene and a laugh, and something else for a companion. It is the bar you go to with friends, not the one you go to alone.

More on this site: the best AI girlfriend apps, free AI chat apps, the best AI companion apps, AI roleplay chat apps, all AI companion app reviews.

A cowboy, a senator, a planner and a tired adjuster arguing over a horse is still a scene I'd cut into a reel. The part I'd skip is the yearly plan that outlives the interest, and I'd tell anyone to avoid it.

If you go in, go in for the party. Pick a strange cast, give them a problem and let them run with it. Keep the photos off, keep the card in the wallet until a character earns it, and leave while it is still funny.