Chinese Forge: a small Estonian company's HSK app with 1,342 installs, worth watching for how it got onto the charts on a small budget

Chinese Forge is a Chinese reading app built for the new HSK (HSK 3.0, nine levels). It launched in late September and had only 1,342 installs worldwide by November 5. It showed up on the education free charts in Thailand and Vietnam, but not because the product is great. Three small things did it: from late October it ran Google app ads with Thai and Vietnamese creatives, it rewrote the store page for each country, and it happened to coincide with HSK's December 13 syllabus switch. The playbook is worth copying; the category itself isn't worth entering.

37 videos no one watches

Around Oct 27, a YouTube channel called Regna Vita Studios uploaded 37 videos in one go. Every title follows the same template: "Chinese Forge Portrait Introduction with Mystic Music (Th)" and "Chinese Forge Square Introduction with Deep Music (Vi)". Each runs 16 to 21 seconds, in portrait, square, and landscape versions, with the background music swapped around, and the language in parentheses at the end: Thai, Vietnamese, Spanish. The channel has 1 subscribers, and every video has single-digit views.

These videos were never meant for people to watch. Google's App Campaigns require advertisers to host video assets on YouTube, and it's best to supply all three aspect ratios, which Google then assembles into ads. Single-digit views suggest they're just an asset library.

5 days later, in November 1, Chinese Forge's installs started climbing: on Google Play they went from 284 to 404, then Nov 2 893, then Nov 5 1,342. Its rank on Thailand's education free chart rose from #142 in Oct 29 to #84, and in Vietnam it moved from outside the chart to #139–#145. The timing matches and so do the languages. That it's running paid ads is an inference (medium-high confidence); we didn't directly find its ads in Google's Ad Transparency Center.

What it is

A Chinese reading app for "serious adult learners". The first line on its website reads: "Master Chinese through context, not by wearing yourself out with flashcards." The store description opens with the "flashcard illusion": you drill 200 cards and feel you know the language, then open a real Chinese article and face a wall of characters.

Its features (website and store page, 11-06):

• 4,200 graded texts across HSK 1 to 9, with audio; the advanced levels include original texts by Confucius, Laozi, Sun Tzu and Li Bai
• Tap any word for a popup with pinyin, definition, etymology, grammar and example sentences, with no need to leave for a dictionary
• A handwriting canvas that recognizes strokes and stroke order offline on the phone
• FSRS spaced repetition (the review algorithm the new Anki version uses), plus a hands-free autoplay for commuting
• Claims to cover all 11,092 words and 572 grammar points of the HSK 3.0 syllabus

The free version only opens the HSK 1–2 texts. The "Scholar" subscription unlocks up to level 9, in monthly, 3-month and yearly plans, with a 7-day free trial, listed at $5.99 to $39.99.

It's published by RegnaVita OÜ, a company registered in Tallinn, Estonia, in September 2025. Its registered phone number is Turkish, and it has only this one app. There is no iOS version.

Why now: the HSK syllabus change

CTI (Chinese Testing International), the body that runs HSK, announced in early September that the new HSK 3.0 levels 1–6 will have their first global test in Dec 13, 2026. The old version had 6 levels and about 5,000 words. The new one has three bands and nine levels, with 11,092 words, under the same name "HSK 4". The new version requires 2,000 words, while the old one required only 1,200. The last session of the old version is Nov 7, and registration for the Dec 13 paper-and-pencil test closes Nov 16. None of the 11 sessions in CTI's 2027 schedule, published in Oct 27, is listed as still using the old version.

In other words, everyone studying for HSK around the world will need to switch word lists, textbooks and apps this winter. Chinese Forge launched on Sep 24, and all its copy says "HSK 3.0" or "HSK 1–9", so it caught this wave.

Where users come from: ads plus a store page rewritten per country

Ads are only half of it. The other half is the store page. The Thai listing title is "Chinese Forge: HSK 1-9 ภาษาจีน" (ภาษาจีน means "Chinese language"), the Vietnamese one is "Chinese Forge: Tiếng Trung HSK", and the descriptions are also in the local language.

The Vietnamese description isn't a straight translation. Its first sentence is written for Vietnamese readers: "Vietnamese speakers have a big advantage in learning Chinese thanks to Sino-Vietnamese pronunciation. It sounds easy, but many people fall into the 'passive recognition trap': they know each card on its own, then open a real Chinese article or an advanced HSK exam and freeze in front of a wall of characters." The Thai and Vietnamese descriptions both state that taps show definitions in Thai and Vietnamese (Google Play Thailand and Vietnam store pages, 11-06).

It also has a website, and the whole page is an FAQ: "What's the difference between the new HSK 3.0 and the old HSK?", "Does it cover HSK 7–9?", "Can I read the original texts of Confucius and Laozi?" These are questions test-takers actually search for, and the page was still being edited as of Nov 2.

Beyond that there are almost no other channels: it doesn't show up in the Meta ad library, nobody mentions it on X, TikTok or Xiaohongshu, the Instagram account has 12 followers, and there's no press coverage. The store listing has no ratings and no reviews yet.

Why does a thousand-odd installs in Thailand and Vietnam get it onto the charts? In November 6 it ranked #80 to #95 on Thailand's education free chart, alongside apps with tens of millions of installs like Quizlet, AnkiDroid and Busuu. Category charts rank by recent install velocity in that country, and the education tier in Thailand has a low threshold. So the rank doesn't reflect scale. It only shows that a small budget was enough in a small category.

Who it's competing with

It's crowded. Google Play installs, 11-06:

Further down is a long tail of indie developers: Wanyu, HanziFeed, ReadMe HSK, HanziForge, Hanzi Mind, all building "HSK 3.0 + leveled reading + FSRS," all very small. Wanyu even went fully free with no ads. 2026 Starting June this year, small apps like Baolingo, Huahua, Nhai HSK and DayPrep also appeared, targeting Vietnam and Indonesia, with installs ranging from a few hundred to over ten thousand.

What Chinese Forge adds beyond the pack is HSK 7–9 plus classical texts, with reading, handwriting and review all in one app. But those selling points are built for advanced learners in the English market, while the Thai and Vietnamese ads bring in mostly beginners (inference). Its current numbers come from ads and localization, not from these differentiators.

Will the money come back

Unknown. In Thailand and Vietnam an install is cheap, but conversion and ARPU are also low. There is no data yet on whether free access to only the 1–2 levels can move beginners into a subscription: no ratings, no reviews, no revenue. It ships a new version almost daily (10-29 from 1.0.33 to 11-06 1.0.37), which looks like building while buying traffic, using cheap markets to test creatives and conversion (inference).

If it were you

C: copy its playbook, not its category. The HSK app space already has tens-of-millions-install incumbents, official-backed apps, a Vietnamese local team and a string of indie developers, and Chinese Forge has yet to prove it makes money itself.

Who it suits: a 1–2 person small team that wants to build an exam-prep app right as an exam changes format, has an exam that isn't yet crowded (JLPT, TOPIK, DELE, IELTS-type exams), and is willing to spend a little on Google app ads.

Who it doesn't suit: anyone wanting to build yet another English-interface HSK reading app (Du Chinese and a pile of indie developers are already there); anyone who only wants to build product and won't touch ads (this model lives or dies on paid acquisition and store pages); anyone targeting Vietnam (local teams' dictionary and exam-prep apps already have millions of installs).

How to do it: watch an exam's new official syllabus, build a content library covering "100% of the new vocabulary list and grammar points," make the beginner level free, unlock all levels with a subscription, and include a 7-day trial. Turn the website into an FAQ, with one question per search term ("What's the difference between the new and old versions?" "Does it cover the superlative?").

How to acquire users: follow its creative matrix. For each language, make vertical, square and landscape formats, each paired with two or three music tracks, 15–20 seconds long, uploaded to YouTube and left to Google's app campaigns to optimize. Rewrite the store title and description per country, and open with the concrete difficulties locals face in learning the language (like its Vietnamese copy about Sino-Vietnamese readings). Pick one country where the category chart bar is low, spend a little, and see whether it reaches the top 100 within a few days, then check whether trial-to-subscription converts (inference).

Deadline: demand for exam-prep apps around a syllabus change clusters around the first sitting of the new version. This HSK wave is already too late to build from scratch (first sitting Dec 13, paper registration closes Nov 16). When the next exam's revision is announced, that's the moment to start, and you should be live at least three months before the new version's first sitting.

Where the read could be wrong: ① By the end of 2026, Chinese Forge had collected dozens of ratings in Thailand and Vietnam and charted in the education top grossing list, suggesting Southeast Asian beginners will pay for advanced reading, so this category is wider than it looks; ② after it ran Spanish creatives, it also surfaced on the Spanish-speaking education charts, suggesting this low-budget approach can be replicated country by country; ③ conversely, if after the Dec 13 first sitting its new installs drop back to a few dozen a day and it stops updating, then it was a test that didn't work, and the only thing left to learn is the creative format.

Sources

Google Play store pages (English, Thai, Vietnamese) with install counts and category chart positions, 2026-10-29 to 11-06; the website chineseforge.regnavita.com (updated 11-02, read 11-06); the YouTube channel @regnavita and three ad creative playlists; Estonian company registration (regia.ee, e-krediidiinfo.ee); HSK 3.0 exam dates: the Shanghai Municipal Education Commission English site (2026-09-15), and exam guides from Khanji School, StudyCLI and StudyU; the HSK GO website hskgo.com; the Meta ad library (11-06, no records). Where we wrote "inference," it is inference.