Repackaging speaking practice as "city quests": Speak City, 57 days after launch, about 100 US-store ratings, just climbed to around the top 100 of the Education chart

Oct 26, a US user named D Dominic gave Speak City a one-star review on the App Store titled "Beware their free trial": "I signed up for the trial to check it out, and got charged on the spot. Tried to get a refund from Apple and was denied." A month or so earlier, another user said it was "too expensive" and also that the free trial still charged their card. This app, launched Sep 10, is AI speaking practice in a game shell: you get "airdropped" into a foreign city, and the coffee shop, pharmacy and taxi are each a level you clear by speaking. As of 11-06, it had 111 US ratings, averaging 4.6, and ranked #113 on the US Education Top Free chart. Below: who makes it, how it's getting traction, how it makes money, who it's competing with, and whether you should chase it.

What it is

An iPhone app, titled "Speak City: Language Game", English-only interface, about 200 MB. The first line of the store page isn't about features, it's about an awkward moment: "You've studied, you've taken the classes, and when a real person says a sentence to you, your mind goes blank."

The loop has four steps: land in a city where nobody speaks your language, with only a passport; each level is a real place with a real goal, like "order breakfast from the barista," "get medicine from a chatty pharmacist," "blend into a party"; you say it in your own words and the character responds; at the end of each level you get a "fair verdict," a transcript of what you said, what was right, what to fix, and you can retry for free. Clearing a level stamps your passport and unlocks a new neighborhood on the city map, and the map itself is your progress bar.

Two lines on the store page are aimed at Duolingo: "AN HONEST VERDICT, NOT A GOLD STAR" and "You can never fail, but you can't fake it." When you get stuck, the character slows down, rephrases, and gives hints, so a total beginner can get through, but the verdict won't lie to you.

Revenue comes from the subscription Speak City Plus, which unlocks the whole city. The US store page lists two tiers: $17.99 and $83.99 (App Store US, 11-06; the page doesn't say which is weekly or annual, but judging by price, the lower one is likely monthly or weekly and the higher one annual).

Who makes it: an app factory in Istanbul

The developer account is Myrtle, and the seller listed is Mer-Tel Elektronik, a company in Istanbul (mertelapps.com). The same company also has Tipple (quit drinking, paired with a pet), Yeti (screen time, also with a pet), Poco (AI pet), Good For You (supplement scanning), Leti and others. The playbook is the same: one narrow AI capability, wrapped in a cute "game" or "pet" shell, with a punchy hook line, each product gets its own Facebook page pushing short videos, and the website is just a one-page landing site. The body text of Speak City's site, languagegame.ai, is two sentences.

Its iteration pace is factory-style too: 7 versions in two months, with 1.0.7 shipping on 11-06, and the package size trimmed by about 16% (from AppInsider's version timeline).

How fast it grew

Not fast, just starting to gain traction. Total US ratings (App Store US):

30 new ratings in the last 7 days. It has never charted on the overall or Top Grossing charts, and globally it has only charted on a US category chart. AppInsider estimates 10k+ global installs (a third-party estimate, unverified), and it was first indexed on Oct 12.

For comparison, LingoLooper, launched two years ago and closest in form, also got +27 US ratings in the last 7 days. Speak City's momentum is about the same as a two-year-old product, and the absolute numbers are small. The bar for the US Education Top Free chart is low, so picking up 5 to 10 extra ratings a day is enough to reach the top 150.

Where users come from: short-video ads on Facebook and Instagram

There's a Facebook page called "Speak City" categorized as a software company, with several videos carrying the same copy: "Your Next Trip Starts Here ✈️", followed by "Order breakfast 🥐. Get a prescription 💊. Take a taxi 🚕. Meet locals 👋", ending with "tap the button below." The same video appears in several posts with different IDs, all with download buttons, which is what ad creative variants look like. A few swap in Spanish and Italian voiceovers, such as "A whole Italian city. Every cafe, every pharmacy is a level." One Reels post by 11-03 had 1.5K views as of 11-06.

The copy never mentions AI, only everyday situations. Breakfast, medicine, taxis: understandable in 3 seconds, perfect for vertical ads. How many ads they run and in which countries isn't visible from outside.

Other channels are thin: the name includes the phrase "Language Game," which ranks for terms like "english language game"; sister apps Tipple and Good For You show it in their "you might also like" sections. Searching "speak city app" on TikTok returns other brands like Speak and Jumpspeak; nobody is organically talking about it on X, Reddit, media or Product Hunt.

Is the money coming in?

Can't tell yet. No placement on the grossing charts suggests revenue is still small. The funnel is the classic "ads → paywall → trial-to-subscription." Among the existing reviews, the trial is the notable part:

• "Signed up for the trial, charged immediately, and Apple won't refund." (D Dominic, one star, 10-26)
• "Too expensive. Looks fun though. Says free trial, but still charged my card when I tapped it." (renae3!, two stars, 9-16)
• "Kept getting stuck, asking the same question over and over. I love how real it feels." (NapoliSun, three stars, 10-04)
• "Immersive speaking is a great idea, but you have to tap several buttons to understand what they said and what you should say. Hints are audio only, no subtitles." (Fondon12, three stars, 9-17)
• "Awesome! I paused for a while because of money, and support was really helpful." (AmbDee, five stars, 11-04)

Two complaints about trial charges, one about price: the factory has tightened the paywall, which usually means it's seriously working on payback math. On the other side, "stuck, repetitive, no subtitles" shows the conversation layer is still rough. 4.6's average rating doesn't quite match these reviews, and how much of the rating is prompted is hard to say.

Who it's competing with

"AI to practice speaking, role-play in real scenarios" has been a red ocean for a while. Speak City's difference is the packaging, not the tech.

Also, "a town that speaks a foreign language" is a common hackathon idea, and web prototypes like Talk Town, Maple Street and LangCity are a dime a dozen. None of these are copying Speak City; they're independent takes on the same idea. Which means the idea is easy to come up with and the barrier to entry is low.

What's driving its traction

Break it down, and quite a lot is available for followers to copy:

1. Pain-point copy: "Studied it but can't speak it," aimed at people who were already taught by Duolingo and feel the gap.
2. Shell that's easy to film: cities, passports, stamps, unlocking neighborhoods, understood at a glance, and easy to turn into Reels. "AI tutor" is already overused by Praktika, Speak and Loora, while "game" makes people more willing to tap download.
3. Reverse positioning: "Honest verdict, no gold stars," giving existing users a reason to switch apps.
4. Knows how to run ads: the same company ran this playbook on Tipple, which added 142 US ratings in 9 days.

What followers can't copy is this factory's ad know-how and cross-promotion between products, but neither is exclusive. The real problem is elsewhere: the product mechanic has no moat, the conversation experience is still rough, and at this volume there's no proof of retention or payback.

If it were you

One-line take: this playbook of "AI capability + game shell + pain-point copy + short-video paid acquisition" is worth learning, but building another Speak City for language learning isn't worth it.

Who it suits: a small team that already runs Meta ads, can produce vertical creative in volume, and has a working AI voice-conversation pipeline.

Who it doesn't suit: anyone hoping to start from organic traffic or SEO. There's no organic content about it on TikTok or X; without paid acquisition there are no users. It also doesn't suit anyone trying to go head-to-head with Duolingo or Speak; their money and reach are out of your reach.

How to do it: don't build "yet another city for learning a foreign language." Move this shell to an adjacent, not-yet-overdone niche of "speaking practice." For example: English job-interview levels (one interviewer per level), sales-call levels, workplace language for healthcare or hospitality workers, the first-week survival scenarios for international students. Keep three things: real scenarios as levels, a visible map or passport as the progress bar, and "honest verdict + transcript of what you said." Do better where it falls short: laggy conversations, hints without subtitles, and the complaints about trial charges. Those are ready-made points of difference.

How to get users: copy its Facebook / Instagram Reels approach, with creative that talks about scenarios, not AI ("The interviewer asks why you left your last job, and your mind goes blank"); put a category keyword in the name for ASO; test on a low-barrier category chart like Education or Business. In the US, picking up 5 to 10 extra ratings a day is enough to reach the top 150.

Is there a window?: no clear window. This direction doesn't rely on a platform loophole, but the "AI speaking game" idea is being built everywhere, and it gets more crowded the longer you wait. If you want to test it, run a small-budget round within the next month or two.

Signs you've misread it: if by December Speak City is still getting only 20 to 30 ratings a week and never ranks on the grossing charts, the packaging is pulling people in but not turning them into revenue, and switching to another category would probably produce the same result. Conversely, if it reaches the top 200 of the US Education grossing chart, or the same company starts pushing it into other countries, the packaging is paying back, and that's when it's worth chasing harder.