
The app isn't live yet, but Douyin already has 350k likes: how msats used "build in public" to chart a fitness app
What it is A Chinese iPhone + Apple Watch fitness app. You lift with your watch on, and the watch auto-detects each completed set and logs it. There's also running with territory capture, calorie estimates from photos, an AI coach, and music lyrics.
Why it took off The developer, June, used Douyin's official hashtag to post "the app I'm building" and built up 46k followers. In November, 2.0 added watch-based set detection, and two ten-second demo videos drove a new wave of attention. Two free months and a promise that core features stay free got it to the Free chart quickly.
CHow you could follow Don't build a Chinese clone to compete head-on. The best thing to learn is its playbook: once the interface existed, it started posting under Douyin's official hashtag, replied "ping me when it's live" in comments, and made each feature its own ten-second video that explains itself. If you have watch development experience, consider the HarmonyOS version, which commenters asked for most.
Est. monthly revenue (App Store) $1K–10KWe estimated this from its chart ranks and rating growth across storefronts, calibrated against verified revenue; typical error is 2–3x. Read it as a tier, not an exact number.
- Data (updated daily) · launched 2026-07-17 · US store rating 1 · charted in 11-20 of 1 stores
iOS · best category chart rank · by store ChinaHong KongMOData 2026-11-08 → 2026-11-20 (13 days, 13 observations). Lower rank is higher; gaps mean it wasn't charted that dayiOS · rating count · US store Number of ratings 1Data day 2026-11-06, only 1 days so far. Apple doesn't give download counts, so rating growth is the proxy
msats is a fitness app for iPhone and Apple Watch built by a Chinese developer. In November 8 it first entered the China-store Health & Fitness Free chart, and the next day it rose to #15. It has only 68 ratings in the store, no press coverage, and nobody has mentioned it on X. Almost all of its users came from the developer's own Douyin account, which had built tens of thousands of "ping me when it's live" followers in the five months before launch. The business doesn't make much money. What's worth learning is the distribution playbook. For anyone thinking of building a similar product, the opportunity is small.
A 33-second video
In 2026 June 27, a Douyin account called "msats" posted a 33-second video titled "vibecoding大赏|Msats app" with one line of copy: "Hi everyone, this is the music-and-sport app we're developing, deeply tying every song to related workouts." The video showed interface demos under Douyin Tech's official hashtag #vibecoding大赏 (per 36 氪, the hashtag drew 38k participants and 1.28B views).
The account had been open for only two weeks, and its first video had 94 likes. This one got 350k likes, 67k shares and 57k saves (Douyin, 11-08). The app wasn't finished, and the comments were a row of "notify me at launch" and "waiting, ping me when the Android version is out." One top comment with over a thousand likes read: "This looks clean. Hope there are no ads when it ships."
Over the next few days, he posted several updates like "thanks for following, going through app review," each getting 10,000 to 30,000 likes. In Jul 9, because the filing took 20 working days, he first released 500 test spots, with people claiming links in the comments (that post has 1,170 comments). The App officially launched on Jul 17. As of Nov 8, the account has 46k followers.
What it is
A fitness app that packs everything in, usable only on iPhone and Apple Watch, requiring iOS 26 or later.
- Strength logging: sets, reps, weight, 1RM, drop sets, supersets. The 5 version in November 2.0 added the main selling point: lift with an Apple Watch, the watch recognizes "set completed," taps done automatically, and starts the rest timer. Detection uses only the watch sensors, no camera. The developer says the model has only about 47k parameters and runs on the wrist.
- Running and cardio: 20 cardio types, 30 running analysis plug-ins, and "Footprint Run": streets you've run get claimed as hexagonal cells that become your territory.
- Other: calorie estimates from photos, an AI coach that builds weekly plans, Apple Music lyric sync, and Msats Focus (you have to finish a set of push-ups before you can open entertainment apps).
- Pricing: most features are free. The in-app purchases on the China store page 11-08 run from ¥3, ¥6 and ¥10 monthly tiers, to Max at ¥90 per year, to Ultra as a one-time lifetime purchase at ¥166. New users get two months of membership free, the auto-detection is publicly promised to stay free, and the Douyin profile even says "I don't recommend paying for membership, try the two free months first."
The developer, named 豪 刘, has only this one app. In November 5 it was renamed "Msats Studio". It's unclear whether this is an individual or a small team. Releases have been frequent, and by November 8 it was already at 2.1.
Why November made it take off
The June wave was read as a "vibe coding project" and didn't directly translate into chart rankings for the app. What actually pushed it onto the chart was three videos in November:
| Date | What the videos cover | Likes / shares / saves (11-08) |
|---|---|---|
| 11-05 | Founding Msats Studio and changing the logo, posted 2.0 | Several dozen likes |
| 11-06 | "Msats' first auto-detection model": finish a set and the watch logs it automatically and starts the rest timer | 2,507 / 1,407 / 1,045 |
| 11-07 | "Cumulative downloads have reached 20k", and says he'll use the feedback from the past two days' 5,000 uses to train a more accurate model | 4,815 / 2,376 / 1,744 |
| 11-08 | "User numbers suddenly surged, the server is under heavy load, upgrading now" | 47 / 18 reviews |
Both videos have over a thousand saves, which means people actually want to install it, not just like the clip.11-08 It first entered the China-store Health & Fitness free chart (#26), 11-09 climbed to #15, and the same day hit #135 on Top Grossing. The China-store rating went from 51 to 68, averaging 4.3 (App Store, 11-08 to 11-09). Hong Kong, Singapore, Malaysia and Macau also charted on the free chart for that category, but none of those stores has a single rating, so volume there is probably tiny. "20k downloads" is the developer's own claim and can't be verified.
It's taking off because a few things lined up:
- The audience was already there. The tens of thousands of people who came in through June's video can see every update now. The November wave doesn't start from zero.
- The new feature is made for video. Flick your wrist and the screen auto-checks the set. You get it in ten seconds. July's "Do 10 push-ups before scrolling Douyin" (19.5k likes) worked the same way: each feature paired with a visual you immediately want.
- Trying it costs almost nothing. Two months free, with core features promised free, so you go from watching a clip to installing without a payment step. The cost is that revenue is thin: it's only #135 on Top Grossing.
- The Chinese-language slot is empty. Xunji, the record-keeping app most China lifters use (11k ratings in the China store), still logs sets by hand. Keep doesn't do automatic set logging on the watch. The closest domestic peer, PROIRON 普力艾, has only 3 ratings in the China App Store.
- It looks good. Reviews like "the UI design is gorgeous" and "it adapts really well to 26" come up a lot, referring to Apple's new Liquid Glass style.
I found no evidence of paid acquisition and no creator reviews. On Xiaohongshu only the official account posts, with single-digit to low-teens likes. Whether they ran DOU+ ads isn't visible from outside.
Is automatic recognition new?
No. Motra (formerly Train Fitness) in the US has been doing the same thing for 2021 years: it uses only Apple Watch sensors to recognize exercises and count reps, claims to support 470 exercises, and has 3,053 ratings in the US store. But it only does strength training, is mostly English, costs $5.99 to $99.99, and has just 102 ratings in the China store. Under msats' 11-06 video, a user with a US IP commented: "I've been using Motra all along, I'm going to try msats."
The Android and Huawei side isn't empty either. Garmin watches count strength reps natively. Reviews have compared rep counting for strength training across the Huawei GT6, Garmin 255, and Coros Pace3. PROIRON is listed on the Huawei watch app store, describing "smart counting, automatic exercise recognition, no manual tapping to finish a set."
msats' version is still rough. Comments under it (Douyin, 11-06 to 11-08):
Only the first exercise auto-logs sets, the rest you still have to do by hand. During rest I flick my wrist and it decides my next set is done. A set should be 12 reps, and when I stop after three or four it still logs a full set.
A Xunji Pro user gave the App Store app 2 stars: "The watch version is still restricted… compared to its headline features like automatic set logging, it basically has no advantage." The system requirement also put some people off: "I want to use it on 18 too, but I tried to download it yesterday and the system version isn't compatible," and "What do I do if I don't want to update to iOS 26?"
So be clear about what this is: automatic recognition is what sparked this wave, but technically it isn't ahead. What it wins on is Chinese language, a lower price, and people talking about it on Douyin every day.
What the comments are asking for
Going through the comments on November's videos, apart from bugs, three requests come up most:
- Android and HarmonyOS: "Can you hurry up with HarmonyOS?" "Please make a HarmonyOS version, there are lots of users" "Will Samsung consider supporting it?" The comments under June's video are full of "Android users, pick me up too."
- Not showing up in overseas stores: "Why can't I find it with a foreign Apple ID?" "Not showing in foreign App Stores." It is searchable in the US store, most likely the search ranking is just too low.
- Will it charge money?: "Does it need payment?" "Does it cost money?" "I hope the main features stay free."
These show two things: more people want to use it than can use it, and what they want is "no money."
Who it's competing with
| Product | Scale | Compared with msats |
|---|---|---|
| Motra | US store 3,053 ratings, China store 102 | The same watch-based auto recognition, earlier and with more exercises; English, pricey, little noise in China |
| Xunji | 11k ratings in the China store, best rank #15 on Top Grossing | Mainstream lifting app in China, lifetime membership ¥88; logs sets by hand |
| Keep | 2.95M ratings in the China store, #1 on Top Grossing | Course platform, does not do automatic set logging on the watch |
| PeakWatch | Nearly 10k ratings in the China store | Focused on recovery and HRV, does not auto-log sets by exercise |
| PROIRON 普力艾 | China store 3 ratings | The closest domestic peer, on iOS and Huawei watches, nobody promotes it |
| Hevy / Strong / Fitbod | 100k to 290k ratings each in the US store | Mainstream manual logging overseas, weak in Chinese |
| INTVL | US store rating 5,694 | Overseas running-community game; msats' footprint run uses the same mechanic |
Rating counts come from each app's App Store page, 11-08 to 11-09.
If it were you
C: The distribution playbook is worth copying, but building another copycat app isn't worth it. msats' real asset is the Douyin audience June built through an official topic, and you can't take that. Its features were already done by Motra and watch makers, and its revenue is thin: two months free, core features promised free, and #135 on Top Grossing. Xunji, the domestic fitness record app that makes the most money, peaks at #15 on Top Grossing.
Who it suits: someone at the 1–2 stage with any app, who can edit short videos and is willing to show up on Douyin and talk about what they're building. What to learn is msats' process, not its features. Another group should think it through: anyone who already has watch-sensor recognition models or Huawei watch development experience, who could build automatic set logging for HarmonyOS. The people asking for it in the comments are real.
Who it doesn't suit: anyone who wants a "Chinese Motra" to go head-to-head with msats. It already owns that positioning and has an audience. Anyone hoping to make money from a fitness record app: users in this category in China have low willingness to pay. Anyone aiming at the English market: Motra has been there for four years.
How to do it: copy its rhythm. Start posting once the app has a working interface. Pin the official topic from the platform on the first video (back then #vibecoding大赏, find whatever the platform is pushing at the time). In the comments, collect "pick me up when it's live," and before launch use test slots to turn follows into first users. After that, make a separate ten-second video for each feature, with a visual that makes people want it immediately (push-ups to unlock Douyin, a wrist flick to auto-check a set). Early on, use the free trial to trade for reach and feedback, and treat user corrections as material for improving the product. For a HarmonyOS version, first look at what PROIRON and the built-in strength counting on Huawei watches already do, and build only the parts they do badly (automatic set grouping, rest periods, a Chinese exercise library).
How to get users: entirely through your own Douyin account. Post "here's what I'm building" first, not "here's what I built." Make each video about one feature. msats posted about a dozen Xiaohongshu posts that almost nobody saw. Short-video platforms are the home turf.
When to move: there is no window for copying the distribution playbook, so you can start any time. For a HarmonyOS version, you need to beat msats shipping its own Android/HarmonyOS version. Given how many comments are asking for it, it will get built eventually, so a working version within six months is the only version worth making.
What would show you read it wrong: ① after msats' two-month free periods expire (the first batch around December), it still holds a top-50 spot on the free Health & Fitness chart and climbs on Top Grossing, which would mean this category monetizes in China better than expected and needs a fresh look; ② it ships a HarmonyOS version before spring 2027, which closes that path; ③ a month of its playbook and official topic and the videos are still stuck at a few hundred likes, which would mean June's video rode a topic that was just starting and the platform's push, something not everyone can repeat.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one Motra (formerly Train Fitness) iOS 2021-03 3,053 ratings in the US store, 102 in the China store Also uses only Apple Watch sensors to recognize exercises and count reps, and launched earlier with more exercises (470+); but strength only, mostly English, priced $5.99–99.99, almost no noise in China Xunji iOS 2019-06 11,191 ratings in the China store, best rank #15 on Health & Fitness Top Grossing The mainstream manual logging app for China lifters, lifetime membership ¥88; no automatic recognition Keep iOS 2015-02 2.95M ratings in the China store, #1 on Health & Fitness Top Grossing Course and content platform, does not do automatic set logging on Apple Watch PeakWatch AI 运动健康助手 iOS 2024-06 9,956 ratings in the China store, best rank #18 on Top Grossing Focused on recovery status and HRV; it records training but does not auto-recognize exercises or auto-log sets PROIRON 普力艾 iOS 2024-12 China store 3 ratings The closest domestic peer. It uses watch sensors to count dumbbell reps, but only within its own plans. Nobody promotes it, and it didn't take off Lianlian Jianshen / Kailian iOS 2023-04 / 2022-02 2,977 / 6,435 ratings in the China store Manual logging plus plans, no automatic recognition Hevy / Strong / Fitbod iOS 2019 / 2011 / 2015 US 102k / 109k / 289k ratings Mainstream manual logging and workout-planning apps overseas, weak Chinese localization; msats is in Chinese, bundles features, and includes automatic recognition INTVL iOS 2024-02 US store rating 5,694 Overseas running-community game; msats' footprint run uses the same mechanic, as just one of its features Gymatic(VimoFit) iOS 2015-10 China store 12 rating One of the first apps to auto-detect exercise reps, and it never took off Apple Watch has built-in "Workout" Big-company features — Every Watch has it Only logs duration, heart rate and calories, no rep counting, no set logging - Where users come from
- Verified Douyin official account + official topic #vibecoding大赏: 06-27 pre-launch demo video with 351K likes / 24K comments / 67K shares; top comment "ship it and tag me" with 74K likes; follow-up videos at 1.3–34K likes; account 11-08 has 46,138 followers (TikHub)
- Verified Douyin feature demo video: 07-24 "Do 10 push-ups before scrolling Douyin", 19.5k likes; 11-06 auto-detection 2,507 likes / 1,406 shares; 11-07 "20k total downloads" 4,790 likes / 2,360 shares, followed by entry into the free chart
- Verified Pre-launch waitlist / TestFlight: 07-09 Since registration takes 20 business days, opening 500 test spots first; grab the link in the comments (1,170 comments)
- Verified Free trial and giveaways: at launch, a QR code in the comments gave one month of Apple Music; new users get two months of membership free from October; auto-detection is promised to stay free; the Douyin bio tells users not to top up yet
- Verified Xiaohongshu official account: a dozen-odd posts, each with single to low-double-digit likes
- Inferred App Store name search: a comment says "found it via dy"; people see the video, then search the store to download (inferred)
- Inferred Paid acquisition / DOU+: no evidence found; TikHub doesn't expose view counts
Why msats (iOS id6779649369) got traction
In one line: msats is an iOS / Apple Watch fitness app by a Chinese indie developer (developer name "Hao Liu", rebranded as "Msats Studio" from 11-05) that puts strength logging, running (with a territory-capture mode), photo-based calorie estimates, an AI coach and Apple Music lyrics into one app. Almost all its users came from the developer's own Douyin account (46k followers): before June was on the store, a demo video showing the interface with the #vibecoding大赏 tag got 350k likes, which collected a waitlist of people ready to download at launch; 11-05 posted 2.0, added "Apple Watch auto-detects each completed set", and the 11-06 and 11-07 demo videos picked up more traffic. The developer said cumulative downloads reached 20k, and the app then charted for the first time in the China Health & Fitness free chart (11-08 #26,11-09, #15).
0. First, correcting the data card
- The card says "rating count went from 3→1→51 in 62 days", which is wrong. 1 is the US-store rating count from the long-tail snapshot (appstore library app_snapshot, us, 11-06); 51 and 62 are China-store figures (cn, 11-08 / 11-09). The actual China-store path is 51 (11-08) → 62 (11-09); Apple's lookup API showed 11-08 late (Pacific time) with 68 ratings and an average of 4.3. These ratings built up since July; they didn't accumulate within the 3-day window. The real new change: 11-08 it hadn't charted anywhere; 11-08 it charted for the first time.
- The prices on the card (mini ¥0.99 / Pro ¥1.99) also don't match the store page. The in-app purchases listed on the China store page 11-08 are: mini ¥3, Pro ¥6, Max ¥10, Max half-year ¥50, Max annual ¥90, Ultra lifetime ¥166 (read via r.jina.ai from apps.apple.com/cn/app/id6779649369).
- Only the China store has ratings. Hong Kong, Singapore, Malaysia and Macau all charted in Health & Fitness free (HK 77, SG 150, MY 187, MO 192), but those storefronts all show rating counts of 0 (appstore library app_day, 11-08/11-09).
1. What it is
- iPhone + Apple Watch, minimum iOS 26, package size 599MB, current version 2.1 (released 2026-11-08). Launched 2026-07-17, in the Health & Fitness category; the developer has only this one app.
- Key features: strength logging (1RM, drop sets, supersets), Apple Watch auto-detection of "set completed" (in-house FSD-Net, about 47k parameters, uses only watch sensors with no camera; the website says 50+ exercises are supported, the Douyin copy says 200+), running and 20 kinds of cardio, 30 running analysis plugins, "Footprint Run" (hex-grid territory capture), photo-based calorie estimates, an AI coach that builds weekly plans, Apple Music lyric sync, and Msats Focus (unlock entertainment apps by doing push-ups).
- Pricing: the free tier covers most features; subscriptions are split into Pro / Max, plus Ultra as a one-time purchase at ¥166. The Douyin bio reads "currently free for two months... I don't recommend paying for membership, try it for two months first." It ranked #135 on the China Health & Fitness Top Grossing chart 11-09 (appstore library chart_rank), which means some revenue, but not much.
- The official site msats.top is just a brand page explaining the auto-detection tech; the download button goes straight to the App Store.
2. Competitors
| Product | Where | Launched | Scale (source) | How msats differs from Motra |
|---|---|---|---|---|
| Motra (formerly Train Fitness) | iOS | 2021-03 | US store 3,053 rating; China store 102 rating (iTunes lookup / appstore library 11-09) | Also relies only on Apple Watch sensors to auto-detect exercises and count reps, and did it earlier, claiming 470+ exercises; but it only covers strength training, is mostly English, costs $5.99–$99.99, and has almost no presence in China (a Douyin commenter wrote "I've been using Motra, planning to try msats") |
| Xunji | iOS | 2019-06 | China store 11,191 rating, best rank #15 on the Health & Fitness Top Grossing chart, lifetime membership ¥88 | The mainstream manual-logging app among China's lifting crowd; no auto-detection. Youxun Pro users tried msats and, frustrated by limits on the watch version, gave it 2 stars |
| Keep | iOS / Android | 2015-02 | 2.95M ratings in the China store, #1 on Health & Fitness Top Grossing | Course and content platform, does not do automatic set logging on Apple Watch |
| PeakWatch AI 运动健康助手 | iOS | 2024-06 | 9,956 ratings in the China store, best rank #18 on Top Grossing | Focused on recovery status, HRV and Watch workout logging; some Douyin creators recommend it for "automatic workout logging", but it doesn't identify exercises or auto-log sets |
| PROIRON 普力艾 | iOS | 2024-12 | The China store has only 3 ratings (lookup) | The closest domestic equivalent: uses Apple Watch sensors to count dumbbell reps, but only works inside its own plans, and it never took off |
| Lianlian Jianshen / Kailian | iOS | 2023 / 2022 | 2,977 / 6,435 ratings in the China store | Manual logging plus plans, no automatic recognition |
| Hevy / Strong / Fitbod | iOS / Android | 2019 / 2011 / 2015 | US store 102k / 109k / 289k ratings; Hevy China store 399 | The mainstream overseas manual-logging type (Fitbod also builds plans); weak in Chinese and localization |
| Gymatic(VimoFit) | iOS | 2015 | China store 12 rating | One of the first apps to do "auto-detect exercises", and it never took off |
| INTVL | iOS / Android | 2024-02 | US store rating 5,694 | The overseas representative of "run a loop and capture the area inside it"; msats' Footprint Run uses the same mechanic |
| Apple Watch has built-in "Workout" | Big-company features | — | Every Watch has it | Only logs duration, heart rate and calories, no rep counting, no set logging |
Takeaway: Motra was already doing "auto-logging strength training with only an Apple Watch" 2021 years ago, so it isn't a new invention. msats is the first in the Chinese market to turn it into something a demo video explains instantly, and it has a ready-made Douyin audience.
3. Where users came from (growth tactics)
All the evidence comes from TikHub's scrape of the official Douyin account "msats" (Douyin ID 57851454525,11-08, 46,138 followers), its video list. The likes, comments, shares, and saves are all 11-08 figures:
| Date | Posts | likes / comments / shares / saves | |
|---|---|---|---|
| 06-15 | "We're developing a fitness, running and music logging app, and we'll take all your feedback" (account opening) | 94 / 13 / 15 | |
| 06-27 | **"vibecoding大赏 | Msats app... the music and fitness app in development" #vibecoding大赏** | 350,765 / 24,331 / 66,693 / 57,108 |
| 06-28 – 07-03 | Follow-up #vibecoding大赏 videos ("thanks for following, now going through the launch process", etc.) | 13.8k, 28.8k, 34k, 12.8k likes | |
| 07-09 | "Filing takes 20 business days, so we opened 500 test slots first; get the link in the comments" | 3,064 / 1,170 | |
| 07-17/18 | Official launch; "If you don't have an Apple Music membership, grab one free month via the comments QR code" | 666 / 131 | |
| 07-24 | "Do 10 push-ups before you scroll Douyin" (promoting Msats Focus) | 19,508 / 398 / 4,141 | |
| 08-09 / 08-13 | Territory run / Footprint Run goes live | 252, 66, 75 likes | |
| 10-07 / 10-09 | UI redesign + two months of free access; the Apple Watch version is free for everyone | 27, 259 likes | |
| 10-08 | "Made the msats MV ad with opus5.5 #Claude" | 42 likes | |
| 11-05 | Msats Studio established, new logo; version 2.0 | 42, 66 likes | |
| 11-06 | "Msats' first auto-detection model... automatically logs a set and automatically enters rest between sets" | 2,507 / 158 / 1,406 / 1,044 | |
| 11-07 | "Cumulative downloads have reached 20k... reinforcement learning based on over 5000 interactions in the last two days" | 4,790 / 167 / 2,360 / 1,735 | |
| 11-08 | "User numbers suddenly surged, the server is under heavy load, upgrading now" | 47 / 18 |
- A Douyin official account is running with the "vibe coding to build an App" topic (verified). 6-27 The video with 350k likes was tagged #vibecoding大赏, an official topic started by @抖音科技. 氪 wrote that the topic has 38k participants and 1.28B cumulative views 36. Back then the App wasn't live yet. The top comments were "Ping me when it's on the store" (74k likes) and "Following now, will there be VIP perks when it launches?" (20k likes). In other words, they built a few tens of thousands of people waiting to download it before it even shipped. Whether the platform gave this topic extra traffic boosts is unverified.
- Waitlist slots before launch and perks at launch (confirmed): 500 test slots claimed from the comments; one month of Apple Music at launch; two months of free membership for new users from October; auto-detection publicly promised to be "free of charge".
- A separate demo video for each new feature (confirmed): push-up-unlocked entertainment apps, territory running, and auto-detection each got their own video. The two traffic spikes (06-27 and 07-24) were both "I want that" visuals. Videos 11-06 and 11-07 each got over a thousand saves, showing people actually wanted to use it, not just tapping like.
- Xiaohongshu official account (confirmed, but small): a dozen-odd posts, each with single to low-double-digit likes. A search also turns up one piggyback post titled "how to download msats", showing people are searching the name.
- App Store search: most people likely saw it on Douyin and then searched the store by name to download (07-24 comment says "found it via dy"). This is inferred.
- Paid acquisition: no evidence found. Whether DOU+ boosting was used can't be determined from the data.
- Media / X / creators: x-radar shows 0 posts mentioning "msats" in the last 90 days; web search found no media coverage; no third-party creator reviews turned up on Douyin either, only the official account talking about itself.
4. Why it took off
- It had an audience before it launched. June used the official #vibecoding大赏 hashtag (the tag itself has platform traffic) to post a UI demo that got 350k likes and 67k reposts. The account went from zero to tens of thousands of followers. Every update since has been shown to that same crowd.
- This November wave was driven by a new feature. 2.0's "Apple Watch auto-logging of sets" is a great fit for a video: flick your wrist and the screen taps "Done" for you. 11-06 and 11-07 videos together have about 7,300 likes, 3,766 shares and 2,779 saves; the developer says downloads have passed 20K; 11-08 hit China store Health & Fitness Top Free #26,11-09 for the first time and climbed to #15, and charted on Top Grossing at #135 at the same time.
- You can try it for almost nothing. New users get two months of membership free. Auto-detection is promised to stay free forever, and the Douyin bio even tells people not to pay yet. So the path from watching a video to installing and trying it has almost no barrier. The cost is very low revenue (#135 on Top Grossing).
- The design looks good. Many comments say "the UI is beautiful", "fits the 26 system well", "minimalist style and details". The minimum OS requirement is set at iOS 26, built in the new Liquid Glass style.
- This spot is empty in the Chinese market. Motra does the same thing, but in China it has only 102 ratings, is all English, and is expensive. Big domestic players like Xunji and Keep have no auto-detection on the watch. The only domestic peer, PROIRON, has just 3 ratings. Chinese users who "have an Apple Watch and lift weights" had no local, actively promoted option until now.
5. What's the differentiation, and is it why it broke out?
- The differentiation has three layers: ① the watch alone judges when a set is finished, and starts the rest timer automatically; ② training, running (including territory capture), diet and music all live in one app, in Chinese; ③ it's cheap: the one-time purchase is only ¥166, and the cheapest subscription is ¥3.
- Is that the reason? Partly. Auto-detection was the direct trigger for this wave, and the demo video and the "20k downloads" post are both built around it. But the deeper cause is distribution: June's Douyin account has 46k followers, built on the vibe-coding topic, and auto-detection just gave that crowd a new reason to download. The feature is still rough, though. Comments say "only the first exercise auto-logs sets, the rest are manual", "flicking your wrist between sets counts as finishing a set", "a set of 12 counts as three or four reps". Version requirements also turned some people away ("I want to use it on 18 too"; one Xunji Pro user gave it 2 stars over watch compatibility). Technically Motra is more mature. msats wins on Chinese language, price, and having people pushing it.
6. How many people can see it
- It's been circulating in small circles. The official Douyin account has 46k followers. June has a hit with 350k likes, but it was watched as a "vibe-coding project", not as a fitness app. No media coverage, no discussion on X, no disclosed revenue, and no TrustMRR listing or founder revenue posts found. Xiaohongshu buzz is very small. Rating counts across overseas stores are 0, and some Douyin comments say "can't find it with a US store or foreign ID" (the US lookup actually works, so it's likely just poor ranking in store search).
- Copycats: none found that specifically copy it. PROIRON and Motra both came earlier, so this isn't a copycat.
7. What others can learn
- Post the making-of as content, inside official hashtags. Platform topics like #vibecoding大赏 come with built-in traffic, and you can start before launch.
- Build a waitlist before launch. Use comment-section engagement ("ping me when it's live") and TestFlight slots to turn follows into day-one downloads.
- For each feature, pick a visual you can demo clearly in 10 seconds. Unlock Douyin with push-ups; auto-log sets with a wrist flick. The feature itself becomes the video material.
- Lower the entry barrier early with "two months free" plus a cheap one-time purchase, to get volume and feedback fast. It also treats every user correction as training data to improve the recognition model.
- Take a feature Motra has already proven overseas, and make it Chinese and cheap, adding things Chinese users care about (NetEase Cloud Music integration, territory capture, calorie recognition).
8. Still unclear
- "20k total downloads" is the developer's own claim on Douyin and can't be verified. The only indirect check is the China-store rating count (68), and the ratio of ratings to downloads is very low, which suggests the app isn't actively asking users to rate it.
- TikHub can't get Douyin play counts (play_count returns 0). I don't know whether 11-06 or 11-07 were boosted with DOU+.
- I found no official documentation on whether #vibecoding大赏 gives participating works extra traffic support.
- Retention and paid conversion are unknown, and it's unclear what happens when the two-month free period ends. Counting from 7-17's launch, the earliest free periods expire around December (users who joined from October also got two months).
- The developer "豪 刘"'s background (solo or small team) is unclear. The studio was only founded 11-05.
- The database only has data from 11-06 onward. 11-08 wasn't on any chart before that, so the daily download trend from 7 to October can't be seen.
Sources
- appstore library (agent-im server DuckDB): app_day, chart_rank, app_snapshot, review, app_profile (11-06 to 11-09)
- Apple lookup: itunes.apple.com/lookup?id=6779649369&country=cn / us (11-08)
- App Store China page (read via r.jina.ai): apps.apple.com/cn/app/msats/id6779649369 (in-app purchase prices, version history)
- Website: msats.top/
- Official Douyin account posts and comments (TikHub douyin app/v3 fetch_user_post_videos, fetch_video_comments, sec_uid MS4wLjABAAAA_ITlSi61hUw3GLoz_MaL0oVGOLoysOQfapeWJJNWNLZLOSEB6Xj9QK_9rarAAbbJ): key posts 7656105562065696753 (06-27), 7666002363207379699 (07-24), 7704993328526223289 (11-06), 7705623732514503857 (11-07)
- Xiaohongshu search (TikHub xiaohongshu app_v2 search_notes, keyword msats)
- 36 36Kr: "Douyin and Kuaishou creators all betting on AI interactive content": m.36kr.com/p/3948788952628359 (#vibecoding showcase 38K participants, 1.28B views)
- Motra: apps.apple.com/us/app/motra-ai-workout-fitness-coach/id15485…; INTVL, PROIRON, Xunji and Gymatic data come from iTunes lookup and the appstore library
- x-radar search for "msats" over the past 90 days: 0 posts