"The evolved version of zeta": a Japanese indie dev turned AI romance chat into an X where everything revolves around you

Japan's AI romance chat is already crowded with big players like zeta and Chacha. buzoon didn't try to out-human them. It turned chat into a fake Twitter: you post, and your oshi and strangers reply, repost, and DM. It started with a TikTok carousel that got 316k views and collected 800 ratings in eight weeks. The scale is small and the revenue is invisible, but the "being seen" hook is real. Below: how it got going, how it makes money, why this format was tried before without taking off, and whether you should follow it.

A carousel that said "zeta users, please stop for a moment"

On 2026 Sep 15, the TikTok account @buzoon_app posted a 10-slide carousel in black text on white. Slide 1 reads "zeta民のそこのあなた👈‼️ ちょっと止まってくれませんか" (Hey you, zeta user 👈‼️ Could you stop for a sec?). The next slides say: "友達とzetaの進化版を作ったので、皆様のご意見がほしいです" (I built a better zeta with a friend and want your feedback); "推しがフォロワーになるSNSです" (It's an SNS where your 推し becomes your follower); then a comparison slide: "zeta は推しと『話す』けど、これは推しが自分の投稿に勝手に返信してきます" (zeta lets you chat with your 推し, but this one has your 推し replying to your posts on their own); and finally "毎週新しい世界を追加する" (adding new worlds every week) and "最大限改善します" (we'll keep improving as much as we can), with a screenshot of App Store search results and "よかったら調べてみてください" (Feel free to check it out).

This image-and-text post got 316K views, 24K likes, and several comments (TikTok, 300 11-06). The most-liked comment was a joke, "がちやめてくれー😂" (1,230 likes), and the second most-liked was "CMありでいいからずっとできるようにしていただきたい" (ads are fine, just let me keep playing forever, 585 likes). Further down there's a string of "Is there no Android version?" "If it comes out for Android, I'll definitely play"; it only has an iPhone version.

The same day, the developer rushed out a browser version at buzoon.app, posting "アプリなしで PC でもスマホでも遊べます" (play on PC or phone without an app), and this post also got 19k views. The App Store reviews from those days show the carousel's effect: "TikTok見て正直そこまで期待はしてなかった。けど結構いい" (Honestly I didn't expect much after seeing the TikTok, but it's pretty good, 9-17) and "TikTokを見て初めてみました!!すげえおもろい!!" (I saw it on TikTok and tried it for the first time!! It's so fun!!, 9-15).

What it actually is

buzoon (バズーン; the name comes from バズる, to go viral) is a Japan-store iPhone app, listed in Sep 12. The developer account rio nagasaki is an individual account, and the copyright line reads "© 2026 buzz". It was called buzz in its first few days, and the Sep 18 release notes say "buzz は buzoon に生まれ変わりました" (buzz has been reborn as buzoon).

How it plays: pick a world (doting, yandere, childhood friends, reverse harem, BL, host club, royal court, idols... the store page says 200+, the website says 950+), register an account on that world's "X", and post. Residents like, reply, quote-repost, and some secretly DM you. Your follower count goes up and down, and the more followers you have, the differently the characters treat you. You can pick your identity (transfer student, younger sister, new manager), peek at the characters' phones, or write your own setting. If you keep it private, you can even use copyrighted characters.

The most direct user comment: "Twitterで有名人になった気分、承認欲求満たされるのでおすすめ" (It feels like being famous on Twitter, my need for recognition is satisfied, recommended, 10-23). Another wrote: "夢小説の中の自分を描いてた頃を思い出して、まさかこんなアプリがでる時代がくるなんて、生きていて良かったです" (It brought back memories of writing myself into fanfiction. I never thought an app like this would exist. Glad I'm alive, 10-13).

This is where it differs from zeta: zeta sells "being answered by your oshi," buzoon sells "being watched by a crowd."

How fast, how big

The curve is front-loaded: the wave from mid-September to mid-October came from that carousel. After that it fell back to natural growth of about five a day, with no spike from paid acquisition visible.

Comparing with peers shows how small it is: zeta has 216k ratings in the Japan store, adds 7,800 a week, and ranks 33 in overall Top Grossing. Chacha has 14.3k ratings. buzoon is two orders of magnitude behind.

How it makes money

It doesn't force you to watch ads. That's what early reviewers praised most ("広告もなく、本当に自分がその世界の住人であるかのような気分" No ads, and it really feels like I'm a resident of that world). But posting uses up "energy," and free energy runs out fast. That's also the most common complaint: "エナジーが減る速度が早すぎて長続きしない" (Energy drains too fast to keep going). Users' attitude is interesting: a whole stream of people asked to trade ads for energy. In the Sep 20 release the developer actually added "watch a video for gems, up to 60 times a day" and removed the ads shown when leaving a world.

There are two ways to pay: gems (45 for ¥200, 195 for ¥800, 485 for ¥1,800) and the buzoon PLUS subscription, tiered by how many actions you can take per day. Store page prices run from ¥600 to ¥30,000 (the top tier is presumably annual). The uses for gems keep growing: sending gifts to your oshi, and the 1.1.0 release's feature where your oshi reads replies aloud, which costs 30 gems each time.

Some pay: "It's fun, and I've been paying for it too..." (9-27), "I'll subscribe once the month rolls over" (10-29). Others balk at the price: "Tried to pay and it was too expensive lol, gave up" (10-10). In Oct 17 someone commented, "There's weekly and yearly, can you add monthly?" In Oct 28, the 1.0.18 version added monthly.

It has never charted in Top Grossing, and its actual revenue is not visible from outside. Judging by the size of 800 ratings, monthly revenue is most likely still in the tens of thousands of yen range (estimate, low confidence).

Twenty versions in seven weeks

The developer's strongest asset is speed. From Sep 13 to Oct 31, the version number went from 1.0.1 to 1.1.0, roughly 20 releases. A few examples:

• 9-18: renamed to buzoon, worlds expanded to 200+
• 9-19: DMs can include images (a comment asked, "写真が貼れないのはどうして⁉️" Why can't I paste photos?)
• 9-20: Google / LINE sign-in; watch videos to earn gems
• 10-17: after closing the app, the oshi who last DMed you shows up on the lock screen and Dynamic Island
• 10-18: invite real friends into the same world
• 10-27: call your oshi, with real-time replies
• 10-28: go on dates with your oshi; monthly plan added
• 10-29: live streaming, with viewers sending chat and tips
• 10-31: your oshi reads replies aloud in their voice

The developer's replies on the App Store show the same attitude: "キャラから画像が届く・時間が経つとキャラからメッセージが来る・プロフィールの自由度を上げる、はまさに今取り組んでいるところです" (Characters sending images, characters messaging you after time passes, more freedom in profiles: we're working on exactly these now). Users report back: "前にレビューで修正してほしい所を何個か書いたのですが、しっかり全部修正されてあって" (I wrote in a review a few things I wanted fixed, and they were all fixed, 10-22).

This is the one real advantage a solo developer has over a big company like zeta.

Where users come from

1. Official TikTok: 2,400 followers, 7 posts, one went viral, the rest got a few thousand to ten thousand views. The approach is "ask for opinions": "推しのスマホの中身、見たいですか?コメントで『見たい』って書いてくれたら開発が早くなります" (Want to see what's on your oshi's phone? Comment "want to see" and development gets faster, 9-21). Hashtags go straight to #zeta #チャチャ #夢女子.
2. A batch of small accounts on TikTok: from mid-October, at least a dozen accounts with zero to a few dozen followers each posted one or two "app review" carousels a day in nearly identical formats. Captions like "愛のある辛口の個人の感想です。buzoon(推しがフォロワーになるSNS)だけは褒めました" (Honest, harsh personal opinions with love. Only buzoon (an SNS where your oshi becomes your follower) got praise), "zeta民注目" (zeta users, take note), all ending with #zeta #chacha #アプリ紹介 #PR. Most single posts got a few hundred views, the top about 4,300. One template got garbled and read "buzoonってお金の管理がゲームになるアプリ" (buzoon is an app where money management becomes a game), a sign of batch generation. Who runs these was not verified. From the rating curve, these posts only sustained the long tail of about five a day.
3. Official X account: under 600 followers, used to collect requests and report bugs, not as a growth channel.
4. Store featured placement (inferred): under "その他のおすすめ" (Other recommendations) in the App Store, it appears alongside Chacha, momomolo, and other peers.
5. Paid acquisition: none found.

No media coverage, revenue undisclosed. People who know it are still mostly limited to the TikTok fandom crowd.

This format has been done before

"A social network where everyone except you is AI" was not invented by buzoon:

So buzoon didn't win on the idea. It won on three things: packaging (not "AI social network" but "your oshi becomes your follower", aimed straight at dream-girl fans), the image-and-text post that used zeta as its target, and shipping updates nonstop for the seven weeks since.

Its weak spots are clear too, and they're spelled out in the reviews: the AI's Japanese sounds like "ChatGPT-style talk" and "I can't tell what it's saying"; characters who are supposed to be your fans keep insisting they're coming to find you; there's no Android version.

Who's following

Since September, a few more have popped up: Mero (9-23, "an AI lover sends you photos every day"), Pureta (10-06), ito (11-01, BL-leaning), bonnon and others. All are small, and none has copied the timeline format and built real traffic. The real threat is zeta or Chacha adding a timeline themselves, since they already have hundreds of thousands of users.

Another signal: buzoon is already listed on global stores, with store pages in English, Korean, Traditional Chinese, Simplified Chinese, French, German and more, across 11 languages. The US store title is "AI Character Chat Sim", and Taiwan's is "AI角色戀愛聊天·互動劇情SNS". But outside Japan, every store has only 0 to 4 ratings. It wants to go global, but hasn't really gone yet.

If it were you

B: worth a test, but be clear about what you're betting on. The selling point, a social feed where your oshi watches you, is backed by an 310k-view post and a pile of genuine positive reviews. Outside Japan, almost nobody makes the romance version. But the app is still small and its revenue is unproven. The format has also been tried by five or six players in Japan that didn't take off, so whether it makes money is still a bet.

Who it suits: 1–2 people who can ship several versions a week; people who know the speech style of dream-girl / otome / BL / self-insert fandom (that crowd can tell in a glance whether you're one of them); people who can make TikTok image posts.

Who it doesn't suit: anyone who wants to head-to-head zeta in Japan (they're two orders of magnitude apart, and buzoon already owns the "upgraded zeta" framing); anyone who can't write natural dialogue and just wants to plug in a model and launch (buzoon gets the most flak for unnatural AI Japanese); anyone counting on a one-off idea rather than continuous shipping.

How to build it: three paths, ranked by value. ① Change language: English-speaking self-insert / x reader fanfic, Traditional Chinese (Taiwan) otome communities, Korea. buzoon has translated but hasn't pushed these, so the window is before it does. ② Fix what it does badly: Android (a row of TikTok commenters keep asking for it), more natural dialogue (better models, character-specific speech habits), and go deeper on the "oshi's phone" peeking mechanic. ③ Monetize its way: no forced ads, watch videos to earn energy, subscriptions tiered by "how many chats per day" with weekly, monthly and yearly options, gems that sell features (voice, gifts, calls), not extra chat sessions.

How to get users: copy the structure of its image post. Slide one names the competitor's users (swap "zeta fans" for the users of the top app in your market). Slide two says in one line how you differ. Slide three shows a timeline screenshot. The last slide says "I'll change anything you suggest." Run the official account on "tell us in the comments what you want, we'll build it first", shipping as you go. Once it takes off, add templated review-style image posts for the long tail, but don't let the templates get sloppy the way it did.

Timing: before 2027 February, by inference. Reasoning: buzoon's multilingual store pages are already done, so an overseas push could come at any time. And zeta or Chacha adding a timeline is a matter of one or two versions.

What would show you got it wrong: ① By the end of 2026, buzoon is still adding around 5 ratings a day and has never charted in Japan's entertainment Top Grossing, which would mean "being watched" is fun but doesn't make money. ② zeta or Chacha ship a similar timeline or fan-interaction feature. ③ buzoon's ratings in the US or Taiwan stores climb into the hundreds within a month, meaning its own overseas push worked and the gap is gone. ④ The top romance-chat app in your market already has a timeline feature and you just missed it.

---

Data and sources: App Store storefront pages and rating counts for Japan and other regions, plus reviews (9-15 to 10-29 of 58 reviews, 11-06 to view); buzoon version release notes (9-13 to 10-31); TikTok @buzoon_app profile, posts and comments, and posts about "推しがフォロワーになるSNS" (11-06); x.com/buzoon_app; buzoon.app; aitter.chatls.com; PR Times (momomolo, 2026-04-17); TechCrunch (SocialAI, 2024-09-17). Anything marked "inferred" is inference.