Japan had no Japanese version of Finch, so three Korean students filled the gap with a capybara

BeCappy is a self-care app where a capybara chats with you, writes your diary, and keeps you company through habit check-ins. It was made by three students at a university in South Korea. It currently ranks only on a small chart in Japan and is still tiny, but the gap it found is clear: Finch has proven in the US that this kind of app retains users and sells subscriptions, yet Japan never had a Japanese version. Below: what it is, how it climbed in Japan, where the money comes from, and whether you should build a version for another language.

A review from Oct 23

On Oct 23, 2026, a Japan-store user left a long five-star review. They said that for a long time they couldn't bring themselves to do anything, kept putting things off, blamed themselves and felt worse, and their routine fell apart. Now they're trying to slowly get healthy habits back, and "this app really helped me a lot." They said the words and diary entries from キャピー (the capybara's name) felt warm, "as if it cherishes even my ordinary day." Lately they tick off daily habits and gradually get their sleep schedule back on track.

Same day, another user said they're "still a kid and can't pay" and thanked the developer for adding the hot spring they'd suggested, though they don't know how to take キャピー there.

These two reviews pretty much sum up the app: a character that gives no advice and just listens, a check-in system that never nags, and pet-raising where you dress it up and take it to a hot spring.

What it is

The protagonist is a capybara named キャピー (Korean: 캐피), living in a slow little room. You can tell it you're tired today, that you finished a small task, or just "came to see you." The store page is blunt about it: it won't rush to give advice, it listens first; it remembers what you talked about last time and picks up from there; and a day later you can read a "your day" entry written from キャピー's point of view. You don't have to write a diary yourself, but there's one to read.

On top of that is a full Finch-style pet-raising loop: mood logs, daily habits, automatic check-ins. Completing them earns 'hay', which buys hats, hanbok, and room themes. The October update also added a course of "thinking exercises" from 90 sessions, short texts, ambient sounds, and white noise.

It was built by Rho ChangJun and two classmates from the software department at Gachon University in South Korea. The GitHub organization name includes SOMA, which appears to be one team in a Korean government-run AI·SW training program (inferred). Jun 29 it launched first in Korea, Jul 26 globally. Three students built the back end, payments, chat encryption, and fixed crisis-topic responses, and much of the code was written with AI coding tools.

How big it really is

On the Japan Health & Fitness free chart, from late October to early November it bounced between #67 and #171, and ranked #109 at 11-06. That chart's bar is low: many apps ranked 50 to 100 have only dozens or hundreds of ratings. Third-party estimates put its lifetime global installs just over 10k (AppInsider, estimate). It has never charted on any grossing list, and its revenue isn't visible from outside; it's likely still very small.

So this isn't a proven hit. It's an early signal: in this one market and this one small category, people are slowly finding it, sticking around, and writing long reviews.

Why Japan

Finch was English-only in Japan. Finch has 770k ratings in the US and only 834 in Japan. It's not that Japanese users don't want it; it just never had a Japanese version. Local peers are either cognitive-behavioral therapy apps with AI diary analysis, like Awarefy and muute, which lean toward "therapy," or chat-only apps like Cotomo with no pet-raising or habits. "Cute character + chat that remembers you + diary written for you + habit pet-raising" was, on Japan's charts, unique to BeCappy.

It was built for Japanese users, not translated over. The Japanese copy is native-level, and the store page's opening lines read like "キャピー、今日はちょっと疲れちゃった…". The free quota is also most generous for Japanese users (about 45,000 tokens a day, versus 40,000 for Korean and 35,000 for English). "Doesn't give advice, just listens" fits a very Japanese need: wanting someone to hear you out without wanting to bother anyone.

What keeps people around is the diary and check-ins, not just the chat. One August user said they're the type who can't stick with ordinary to-do apps. BeCappy's habits are "optional, but doing them earns hay and an animation, like getting a little praise." The next-day diary gives a reason to look in every day.

Where users come from

• Organic store search and category charts. Confirmed, and the main driver. In October 18 they renamed the app separately for each language: the English name went from the slogan "Don't Worry, Be Cappy" to the category term "Self-care Capybara", and the Korean name added 셀프케어 카피바라 (self-care capybara). Screenshots went from 6 to 9. Reaching #100 on the small Japan chart also brought more browsing.
• Rating prompts only for heavy users. Confirmed. Only people past a cumulative chat volume threshold (around 15,000 tokens in Japan and Korea) get asked to rate, so those asked already like it, which is why the ratings come in at 4.7 stars. Each major release triggers a prompt.
• Official Instagram. Confirmed. Short videos started in September on three accounts (Japanese, English, Korean), with lines like "when you don't feel like working, use BeCappy" and "ask キャピー about me", tagged #adhd #studywithme. Japanese creator videos tagged #pr have also been spotted, though I couldn't verify specific accounts.
• TikTok. The official account has 10 followers, with single videos getting a hundred to a couple thousand views. Negligible.
• Paid acquisition. Some preparation was done: in July the privacy policy added Meta ad measurement, and in August it connected Meta's install attribution. On Nov 1 a daily report with "Meta spend / installs / CPI" was added to the back end, then removed the same day, with the submission note reading "cancelled campaign." No active Meta ads appear in the ad library. It looks like they tested or prepared for paid acquisition without running it sustainably (inferred).

No media coverage, and no one mentions it on X.

Where the money comes from

From 7 to September it was a "launch free period" where everyone could use it freely. Sep 23 it added subscriptions, mid-October it cancelled the 48-hour free trial for new users, and on Oct 20 it cut free chat to roughly 10 rounds a day. The shift is from pulling users in to making money.

Both price tiers are listed at once, which looks like price testing. Subscribers get more than ten times the chat quota, and first-time subscribers get hay. There are also banner ads, and watching ads earns hay. One October user wrote, "if it's 1000 yen a month, I'd pay if I can stick with it": willing to pay, but still on the fence.

On the cost side: the core is a large-model chat companion, so every heavy chatter adds model costs, and the free-quota line is where the margin is decided.

Who it's competing with

Between 2026 and 9 November, a batch of Finch-style pet self-care clones also appeared (Mipkin, Poppets, Soffi, MeowCare; Japan has ことり and めばな), each with 0 to 6 ratings. In other words, "cute animal + check-ins" already has plenty of players, and the capybara itself isn't the barrier. What BeCappy adds is the chat companion that remembers, the next-day diary, and real localization.

If it were you

The verdict: What's worth learning isn't the capybara; it's the formula. Take Finch, which the US has validated, localize it into a language it hasn't properly served, then layer on what large models can do and the original lacks: a chat companion that remembers you and a diary written for you. BeCappy has shown in Japan that this formula keeps people around and gets long reviews, but it hasn't yet shown how much money it can make.

Who it suits: A one- or two-person team whose native language is one Finch hasn't seriously served (Traditional Chinese, German, French, Portuguese, Indonesian, Thai, etc.), who can write warm copy and make cute characters and small animations; or a team that already has an AI chat backend and wants a lower-pressure, stickier shell.

Who it's not for: anyone who only does English and wants to go head-to-head with Finch in the US; anyone who can't draw a character people want to dress up; anyone not ready for the responsibility of mental health content (self-harm topics, minors). These apps have kids using them. BeCappy only added a fixed crisis response in November.

How to do it:

• Pick a big market where Finch only exists in English. First check whether the local competitors all lean toward 'therapy / analysis'. The gap is on the 'cute companion' side.
• Use an animal and a tone locals already like, write the copy in the native language, don't translate it, and give example opening lines so the first message is easy to send.
• The next-day character diary is a must. It's what sets it apart from every reskinned pet app, and it's what users mention most in reviews.
• Get reviews fast with a free phase, then tighten up like it did: cancel the trial, about 10 free rounds, subscription quota ten times higher. Set quotas by language and tune that line with real data.
• Only show the rating prompt to people who chat a lot.

How to get users: start with organic store traffic. Name the app with the category words locals search (self-care, diary, mood). Write the subtitle and screenshots separately for each language. Once it ranks in a small category chart, rankings bring in their own traffic. For content, follow its Instagram playbook but do it more seriously: short videos of companionship moments like 'when you don't want to work' or 'say one thing to it before bed,' tagged with study, ADHD, and self-care. Then find a few local mid-size creators for #pr tests.

When the window closes: within six months. Chatbot costs have dropped a lot, the Finch-reskin wave has already started, and someone will add AI chat next. If Finch ever releases a Japanese version or adds AI itself, this gap gets smaller.

What would show this read is wrong: if BeCappy still can't hold the top 100 in Japan's category chart by December and never breaks into the grossing charts, the system keeps users but doesn't collect money, and model costs eat it. On the other hand, if it reaches the top 50 of Japan's health & fitness grossing chart, or Korea and English markets take off too, the formula is more general than this analysis assumes, and it's worth moving faster.