
A Tokyo founder who plays pickleball five days a week built an AI match-analysis app that "works even in a gym": three days after launch, 11 ratings
What it is: Film a doubles match on your phone and upload it. The app cuts out dead time, gives you ball speed, shot types and win/loss stats, and an AI chat tells you what to work on. You mark the four corners of the court, so it works in a gym covered with other sport lines.
Why it's getting attention: It isn't actually popular yet. A press release picked up by media outlets, plus the founder announcing it on Threads and X, brought in 10 5-star ratings in one day, two days after the release. Its best position on Japan's Sports Free chart was #113. The bet behind it is that Japan's pickleball player base grows about 7x a year.
CHow you could follow it Don't copy the Japan pickleball analysis angle directly. If you're a heavy player of a sport that's growing, take an AI analysis tool already proven overseas and adapt it to local courts, languages or filming conditions, then cold-start it with a press release and your playing circle.
Est. monthly revenue (App Store) under $1KWe 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) · first listed 2026-11-12 · Japan rating 13 · charted in 11-20 of 1 stores
iOS · best category chart rank · by store Data 2026-11-17 → 2026-11-20 (4 days, 4 observations). Lower rank is higher; gaps mean it wasn't charted that dayiOS · rating count · Japan Data 2026-11-17 → 2026-11-20 (4 days, 4 observations). Apple doesn't publish download counts; rating growth is a proxy for downloads.
On the morning of Nov 17, year 2026, at 9 a.m., a small Tokyo Shibuya company with only 1 insured employees issued a press release announcing PickleballMe, an AI video analysis app for pickleball. Its selling point is something specific to Japan: most people play in gyms marked with badminton and basketball lines, and overseas analysis tools can't recognize that setup. Three days later it had 11 5-star ratings in the Japan store and reached #113 on the Sports Free chart. It hasn't broken out yet. The people downloading it are mostly the founder's playing buddies. What's worth watching is the bet: Japan's pickleball player base grew about 7x in a year, and it wants to be the first "made in Japan" option. But the gap is narrower than it looks.
An "ご報告" post on Threads
On the day of the release, founder Matsuura Haruki posted a "【ご報告】" on Threads: "For about the past year, I've become completely hooked on pickleball and have been practicing almost every day..." with a link to PR TIMES below. The post had about 2,100 views and 61 likes over two days (Threads, 11-19 reads).
The comments read like a gym group chat. One player wrote "こんなアプリ待ってました!!" (been waiting for an app like this!!). Another said "Saw it on Threads, so happy", and Matsuura replied: "Long time no see! We practiced together in Ikebukuro once." Someone else commented "Already installed it!"
Matsuura was born in Osaka in 1986, graduated from Meiji University, ran a marketing consulting firm, joined fintech education startup ABCash Technologies in 2018, eventually became a managing director there, and left in 2023 to found DWZ Inc. In the press release he says he plays "about 5 days a week", and he developed the app with "a working engineer from a University of Tokyo AI lab". The company's registered social insurance headcount is 1 (compalyze).
What it is
Mount your phone on a tripod, film a doubles match, and upload it. The app automatically cuts out ball-collecting and waiting time, keeping only the rallies. It gives the speed, shot type (drive, net shot, drop, reset), win rate, error rate and placement distribution of every shot, and the AI summarizes the match in two columns: "GOOD" and "Issues". There's also an AI chat box where you can ask things like "give me three things to improve" or "what's my pattern for losing points?" Answers include specific moments, and tapping one jumps to that clip (official site). Currently only doubles is supported.
Its approach to "multi-line gym courts" is simple: after upload, the user taps the four court corners in order (near-left, near-right, far-right, far-left), then marks 4 players (self, partner, and two opponents). Having people mark the court manually sidesteps the hard problem of automatically finding the court among clutter. Technically, a lightweight detection model trained for this runs on the phone to track the ball and people, then sends low-resolution short clips to an external multimodal LLM to judge rallies and shot types. Clips are deleted after about 48 hour, and the raw video never leaves the phone (official FAQ and store description).
Pricing (tax included, official site):
| Tier | Monthly | Yearly | Main quota |
|---|---|---|---|
| Free | 0 | — | 60 minutes of uploads per month (about 3–4 matches), only the last 1 months viewable |
| Lite | ¥980 | ¥9,800 | 150 minutes per month (about 10 matches), all videos viewable |
| Standard | ¥2,980 | ¥29,800 | More quota |
| Pro | ¥9,980 | ¥98,000 | More quota |
Launch promo: subscribe to Lite in the app before Nov 30 and get the first 2 months free, then it auto-renews monthly. iOS only; Android is "in development". 11-18 has since updated to 1.0.5.
Who gave the 11 ratings in three days
| Date | Total ratings in the Japan store | Added that day |
|---|---|---|
| 11-17 | 1 | — |
| 11-18 | 1 | 0 |
| 11-19 | 11 | +10 |
App Store Japan, 11-19. It only charted in Japan, peaking at #113 on Sports Free. It's not on the overall or Top Grossing charts. All 11 ratings are 5 stars, with no written reviews.
Making #100 on Japan's Sports Free chart is a low bar. 10 ratings landed on the same day, about 48 hours after the release, matching the playing buddies in the Threads comments above. It looks more like friends cheering it on, or a rating prompt inside the app, than strangers downloading it organically (inferred). In the days after launch, searching "ピックルボール" in the Japan store didn't show it in the top 24 results, so store search isn't working for it yet.
Where users come from
So far there's been only one wave:
- Press release and pickups: The PR TIMES headline locks in the selling point: "国内競技人口1年で約7倍のピックルボールに、体育館でも使える専用AI解析アプリ". Jiji Press, Impress AI Watch, ASCII.jp and AppBank all picked it up, and PR TIMES Sports, AI Watch and AppBank each reposted it on X. Media like the "uniquely Japanese" angle.
- The founder's own accounts: a pinned X announcement, bio changed to "ピックルアプリ創業者" (PickleApp founder); the Threads post at about 2,100 views; also posted on Instagram.
- 2 months free: used to pull in trials. Real paid conversion won't be visible until the promo ends at the end of January next year.
The press release says he'll later sponsor tournaments and support pro players, but no action on that yet. No TikTok, creator partnerships or paid ads found.
How wide is the gap in Japan?
Products like this were already being built in the US three years ago, so the features are proven. The question is how much room is left in Japan:
| Product | Form | Japan's situation | Compared with PickleballMe |
|---|---|---|---|
| SwingVision | iOS / Android, started with tennis, added a pickleball version 2023-12 | Its Japan store name is "SwingVision: テニス&ピックルボール", supports Japanese, 209 ratings in the Japan store; 5,013 in the US store | Real-time ball tracking and 3D ball paths, a large tennis user base; designed for dedicated courts, how it performs in multi-line gyms is unverified |
| PB Vision | iOS / web, 2023-10, about $1.15M raised | English only, 0 ratings in the Japan store | Features almost one-to-one (dead-time removal, shot stats, AI coach, highlights), plus technical grading |
| PickleRise | Web, by Japanese company Oasis Vista | Japanese first, no app install needed | The most direct rival: also uses phone video for AI diagnosis, but bundles court search, racket guides, a "favorite character encyclopedia", and coaches drawing arrows on the video for feedback. The homepage shows new features being added every week. |
| Sansan pickleball court Ikebukuro × Fujitsu G-Search × AllClip | Fixed venue cameras | 2026-07 started validation, rolling out gradually from September | A big-company project using skeleton recognition plus technical grading, only usable at this venue |
| Other Japanese pickleball apps | iOS | Finding partners, courts and scoring; most Japan-store ratings are 0–13 | No video analysis |
So "Japanese + pickleball AI analysis" isn't empty ground: SwingVision already has a Japanese version and over two hundred Japan ratings, and PickleRise does the same thing on the web and updates often. What PickleballMe uniquely has is "native iOS + four-corner marking for gym use". The four-corner method is easy for anyone to copy.
The market it's betting on
A press release cites Japan's pickleball player count rising from about 45k to about 330k, with roughly 11.9M potential players (Pickleball One Co., Ltd., "Japan Pickleball Market Survey 2026", 2026-430k-person survey). The survey comes from an industry company, so its numbers lean optimistic (inference), but the direction holds: Mizuno 2026-08 has entered the Japanese pickleball market, Sansan opened a dedicated venue in Ikebukuro, and December Ariake hosts international tournaments. More people are playing, coaches and courts are scarce, and the demand for "how to get better" is being amplified.
The pricing is aimed at serious players. The cheapest tier is ¥980 a month, about the cost of a video streaming plan. Pro is ¥9,980 a month, close to the price of a private coaching lesson (inferred). Japan's pickleball players skew toward mid-to-high incomes and older age groups, so this price point could hold up, but no data yet proves it.
If it were you
In short: don't copy PickleballMe's Japanese pickleball AI analysis. Three players are already crowding this spot, and PickleballMe itself hasn't been validated by the market. What's worth learning is the approach: take a vertical AI tool already proven overseas, tweak it for one specific local scene, and cold-start with a press release and a community.
Who it suits: someone who is a serious player of a sport that's picking up, has a circle of playing buddies, and can assemble the product with AI coding tools and off-the-shelf vision models. Ideally, in your country or sport, foreign tools have a clear fit problem: courts, rules, language, filming conditions.
Who it doesn't suit: developers who don't play and just want to pick a trending sport for an app. Cold-starting relies entirely on the community, and you won't get in. If you want to do pickleball in Japan, SwingVision already has a Japanese version and PickleRise is updating weekly. If you want to win on accuracy, video analysis accuracy is a bottomless pit, and big players and venue cameras sit at the other end.
How to build it: first find the one tweak. PickleballMe's tweak is "multi-line gym courts," and its solution is to have users tap the four corners and four players instead of forcing fully automatic recognition. The phone runs a lightweight detection model to track the ball and players, while a multimodal large model judges rallies and shot types. One or two people can ship it in a few months. The free tier gets a quota of 3–4 sessions, and paid tiers are priced by upload minutes.
How to get users: write the press release headline around that tweak and a growth number ("up 7x in a year," "works in gyms too"), pitch it to local press release platforms, and get tech and sports media to pick it up. The founder announces from their own account, and playing buddies pick it up in the comments. Then spread it through local tournaments, clubs, and coaches. The most natural sharing for this kind of tool is highlight clips, so make "one-tap highlights for social media" work smoothly.
Is there a window?: no urgent one. Japanese pickleball is still growing, but once the hype fades and SwingVision fills in the gym scenario, the case for a local small app disappears. If you want to do another sport or another country, the earlier the better.
Signs you've got it wrong: ① after the free period ends at the end of January in year 2027 (2 months), PickleballMe's Japan rating stalls at a few dozen and it doesn't chart in Top Grossing, which would mean serious players won't pay monthly for analysis of phone-filmed footage; ② conversely, if its rating keeps rising and it charts in Sports Top Grossing, the "one local tweak" route is working, and it's worth finding another sport with similar local differences; ③ if SwingVision or PickleRise adds dedicated multi-line gym court support, PickleballMe's only selling point is gone.
Data and sources
- App Store Japan ratings and charts, 11-17 to 11-19; Japan store lookup (ratings and supported languages for PickleballMe, SwingVision, PB Vision) and search results for "ピックルボール", mid-November; US store ratings, 11-19.
- PR TIMES press release, 2026-11-17 09:00; pickleball-me.com homepage, pricing and FAQ; App Store listing description.
- Matsuura Atsuki's Threads posts and comments, read 11-19; X @_atsukimatsuura profile.
- Company headcount: compalyze; PickleRise: picklerise.com homepage; Sansan: 2026-07-10 announcement; PB Vision funding: CB Insights.
- Unknown: real downloads, paid conversion and renewal rates, how many of the 11 ratings come from strangers, actual recognition accuracy in gym settings, and PickleRise's user base.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one SwingVision iOS 2016 (tennis); 2023-12 launched a pickleball edition US iOS, 5,013 ratings, 7 days +11 Started in tennis, multi-sport, real-time line calls and 3D ball paths; designed for US courts, mainly English PB Vision iOS 2023-10 US iOS, 143 ratings; funding about $1.15M Nearly identical features (dead-ball removal, stats by shot type, skill ratings, AI Coach), built for US pickleball, with no Japanese and no multi-line gym court support PickleRise (Oasis Vista Co., Ltd.) Web 2026 (unverified) Unknown The most direct rival in Japan: web-only, no app install. Beyond AI diagnosis it bundles court search, racket guides, an oshi character guide, coach annotations Sansan × G-Search × AllClip AI feedback service Big-company features 2026-07 started validation, rolling out gradually from September Sansan Pickleball Court Ikebukuro (3 indoor and 4 outdoor courts) Uses fixed venue cameras with skeleton recognition, only usable at that venue, not aimed at personal phones Coach Pickle / PickleAI / My Pickleball Coach(PlaySight) iOS 2023–2026 Small US market's AI training plans, coaches and venue cameras, with no Japanese or Japanese court adaptation - Where users come from
- Verified PR TIMES press release and media pickups: published 11-17; reposted by Jiji, Impress AI Watch, ASCII.jp, AppBank, play-pickleball.jp; @PRTIMES_SPORTS, @aiwatch_jp, @appbank reshared it on X
- Verified Founder's personal social accounts: X @_atsukimatsuura pinned the announcement and changed the bio to "ピックルアプリ創業者" (PickleApp founder); the same Threads post got about 2.1K views; also posted on Instagram
- Verified Launch promo: subscribe to Lite (¥11-30 a month) before 980 and get 2 months free
- Inferred Offline playing circle: the founder says they play 5 days a week and often join tournaments; says they'll sponsor tournaments and support pro players in the future
- Inferred Store search: almost no comparable pickleball video analysis apps in Japan's store, search ranking not measured
PickleballMe (ピックルボールミー): a Japan-built AI video analysis app for pickleball matches
Bottom line: it hasn't broken out yet. The early signal is 10 ratings in its first day or so (launch day 2026-11-19), peaking at #113 on the Japan Sports Top Free chart. That's too little to mean much. What's behind it is a Tokyo company of 1 people that put out a press release on PR TIMES 11-17 and announced it on the founder's own X / Threads / Instagram, which brought in the first wave of friends and players. The real point is the bet: Japan's pickleball player count grew roughly 7x in a year, while overseas tools like SwingVision and PB Vision don't work well in Japanese gyms. This one moved first to build a "made for Japan" version.
1. What it is
- Film a pickleball doubles match on your phone and upload it. The app automatically cuts out dead time and keeps only the rallies. It calculates ball speed, point win rate, error rate, placement distribution and shot types. An AI gives a "GOOD / areas to work on" summary, plus an AI chat box whose answers cite specific moments; tap one to jump to that frame. Right now it only analyzes doubles.
- Key differentiator: "マルチラインコート対応" (multi-line court support). In Japan, courts are often shared gym or tennis courts with badminton and basketball lines painted on the floor at the same time. Its workaround is to have users tap the four corners of the court and the 4 players themselves, using this manual calibration step to sidestep the hard problem of automatic court detection.
- Pricing (official site pickleball-me.com, PR TIMES): Free (60 upload minutes per month) / Lite ¥980 a month (annual ¥9,800) / Standard ¥2,980 a month (annual ¥29,800) / Pro ¥9,980 a month (annual ¥98,000). Launch promo: subscribe to Lite before 11-30 and get 2 months free; payment must be linked in the app.
- Privacy approach: raw video stays on the phone, and the server only stores the parsed numbers. When judging rallies and shots, short low-resolution clips are sent to an external AI service and deleted after about 48 hours (official FAQ). In other words, it runs a lightweight detection model on-device and uses a multimodal large model to judge rallies and shot types.
- Developer: DWZ Co., Ltd. (Shibuya, Tokyo), established 2025-02, with 1 people on social insurance (compalyze). Representative: Matsuura Atsuki, a Meiji University graduate who has run a marketing consulting firm and served as a director at ABCash Technologies. DWZ's original business was outsourced sales for online course and coding schools. Development was done with a "full-time engineer from the University of Tokyo AI lab." Only iOS so far, Android in development. The App Store listing date is 2026-11-12; the official public launch announced is 11-17 09:00,11-18, and it has been updated to 1.0.5.
2. Data (our appstore library, as of 2026-11-19)
| Date | Total ratings in the Japan store | New today |
|---|---|---|
| 11-17 | 1 | — |
| 11-18 | 1 | 0 |
| 11-19 | 11 | +10 |
- Charted only in Japan: best rank #113 in Sports Top Free; not on the overall or Top Grossing charts. iTunes lookup shows 11 ratings, all 5 stars; no written reviews in the store.
- Ranking #100+ on Japan's Sports Top Free chart is a low bar. 10 ratings clustered on the same day, all 5 stars, looks more like the founder's circle and playing buddies cheering it on, or an in-app rating prompt, than organic downloads.
3. Competitors
| Product | Where | Launched | Scale | How it differs from PickleballMe |
|---|---|---|---|---|
| SwingVision | iOS / Android | 2016 (tennis); 2023-12 launched a pickleball edition | US iOS 5,013 ratings, 7 days +11 (appstore library); self-described as "#1 tennis & pickleball app" | Started in tennis, supports tennis, pickleball and padel; real-time line calls, 3D ball paths; designed for US courts; mainly English, Japanese support unverified |
| PB Vision | iOS / web | 2023-10 | US iOS 143 ratings; funding about $1.15M (CB Insights / Caplight) | Built for US pickleball, with skill ratings, AI Coach and highlights; first session free; the product form is nearly identical to PickleballMe, but no Japanese and no multi-line gym court support |
| PickleRise | Web (Japan, Oasis Vista Co., Ltd.) | 2026 (exact date not verified) | Unknown | The most direct rival in Japan: same model of uploading phone video for AI diagnosis, web-only with no app install, and bundled with court search, racket guides, an oshi character guide, coach annotations, a consultation room, with features updated weekly |
| Sansan × Fujitsu G-Search × AllClip | Venue side (Sansan Pickleball Court Ikebukuro) | Began validation in 2026-07, gradually opened to general users from September | Big-company project | Uses fixed venue cameras with skeleton recognition for motion analysis and skill ratings; only usable at that venue, not aimed at personal phones |
| Coach Pickle / PickleAI / My Pickleball Coach (PlaySight) / DUPR AI (2026-year plan) | iOS / web / venues | 2023–2026 | Each is small | US market AI coaches, training plans and venue cameras; rating platforms like DUPR are also moving into video AI |
| VantageRally | Web (Japanese personal project) | 2026 | Very small (YouTube views in the dozens) | A personal pickleball video AI analysis project; impact negligible |
| Other Japanese pickleball apps (ピックルなう, ピックルボールベア, ピックルボールスコア, etc.) | iOS | 2026 | Japan store ratings 0–7 | Focused on finding playing partners, courts and scorekeeping, not video analysis |
Verdict: This category was already validated in the US three years ago. Japan's gap is narrow: it needs Japanese, it has to work in gyms, and it needs a local team. PickleRise (web) and Sansan (venues) already occupy that spot. PickleballMe is the first to ship a native iOS app focused on multi-court gyms, but not the first with AI analysis for pickleball in Japanese.
4. Where users came from (growth tactics)
| Channel | Evidence | Confirmed? |
|---|---|---|
| PR TIMES press release + reposts | Published 11-17 09:00, headline "Domestic competitive players up roughly 7x in 1 years…"; picked up by jiji.com, Impress AI Watch, ASCII.jp, AppBank, and play-pickleball.jp; reshared on X by @PRTIMES_SPORTS, @aiwatch_jp, and @appbank | Confirmed |
| Founder's personal social accounts | Pinned announcement on X @_atsukimatsuura (bio changed to 「ピックルアプリ創業者」); the same post on Threads got about 2.1K views (11-19 scraped); also posted on Instagram | Confirmed |
| Offline pickleball circles | The founder describes himself as a 「ガチ勢」 who 「plays 5 days a week and often enters tournaments」; he also said he will sponsor tournaments and support pro players going forward | Partly confirmed (intent shown, no concrete actions seen yet) |
| 2 months free after launch | Pre-order Lite before 11-30 and get 2 months free, to drive trials and build paid conversion | Confirmed |
| Store search | Searching 「ピックルボール」 in the Japan store turns up almost no direct competitors, so search traffic is possible; actual rankings weren't measured (iTunes search API was rate-limited) | Not verified |
| Paid ads / TikTok / creators | None found | Nothing found |
11-19 brought in 10 ratings in one day, about 48 hours after the press release went out, which matches the traffic from PR plus the founder's network.
5. Why (so far)
- Timing: Japan's pickleball population has grown from about 45k to about 330k (株式会社ピックルボールワン《日本のピックルボール市場調査2026》, cited in the PR); Mizuno 2026-08 entered the Japanese pickleball market, Sansan opened Tokyo's biggest dedicated venue, and the PJ tournament and PPA Asia December were held at Ariake. Participation is growing fast, but coaches and courts are scarce, which amplifies demand for 「how to get better」.
- A small local difference: foreign tools assume dedicated courts, while Japanese players mostly play in gyms; users can mark the four corners themselves and use it on multi-court layouts. This is the reason the PR makes clearest.
- Distribution rests entirely on one PR push plus founder networking: no second channel is visible, so ratings came in as a one-day spike, not a steady climb.
- Pricing isn't cheap: the lowest tier is ¥980 a month, the highest ¥9,980 a month, aimed at heavy enthusiasts. Japanese pickleball players skew older and higher-income, so this price could work, but no data yet proves it.
6. Differentiation, and whether it's really why it grew
- Differentiation = 「Japanese + native iOS + multi-court gym use + AI chat that jumps to moments」. Functionally it maps almost one-to-one onto PB Vision (dead-ball removal, stroke stats, AI coach, highlights); the only real differences are localization and the gym setting.
- This differentiation earned media coverage (outlets like to write 「uniquely Japanese」), but it isn't yet why it's growing, because it hasn't grown yet. The moat is also shallow: manually marking four corners is easy to copy; PickleRise already offers Japanese plus web, and Sansan has a big company's data and venues behind it.
7. How many people know about it
- Few people know about it yet: revenue isn't public, it isn't on TrustMRR, and big X accounts haven't discussed it (no relevant posts in x-radar's local corpus over 60 days); there's only one round of press release reposts plus the founder's own accounts. No copycats spotted.
8. What to borrow
- Take a vertical AI tool already validated overseas into a local market growing fast but where overseas products don't fit, use a concrete local scenario (multi-court gyms) as the selling point, and put it in the PR headline.
- Use manual calibration (marking corners and players) instead of hard fully automatic recognition, so a small team can launch within months; combine on-device lightweight detection with cloud multimodal large models to cut development and inference costs.
- A press release plus founder networking gets media reposts and the first batch of ratings, so cold-start cost is low.
9. Still unclear
- Real downloads, paid conversion rate, and renewal rate after the launch promo ends.
- How many of the 10 5-star ratings come from organic users.
- Actual recognition accuracy in gym settings (no user reviews on this yet).
- PickleRise's scale and launch timing; when Sansan's service opens to the public and how it prices.
- Whether later channels include tournament sponsorships, court partnerships, or creators; whether TikTok and Instagram ads are running (not found).
Sources
- App Store lookup:itunes.apple.com/lookup?id=6802977767&country=jp
- PR TIMES 2026-11-17:prtimes.jp/main/html/rd/p/000000002.000193097.html
- Website: www.pickleball-me.com/
- Jiji Press repost: www.jiji.com/jc/article?k=000000002.000193097&g=prt
- AI Watch:ai.watch.impress.co.jp/docs/news/2156913.html ;ASCII:ascii.jp/elem/000/004/452/4452357/
- AppBank X:x.com/appbank/status/2122490514435952877
- Founder on X: x.com/_atsukimatsuura ; Threads: www.threads.com/@atsuki.matsuura/post/Df0xnbTn3VR
- Company info: compalyze.co.jp/company/9011001167711
- PickleRise:picklerise.com/
- Sansan AI feedback service validation: jp.corp-sansan.com/news/2026/0710_03.html
- PB Vision funding: www.cbinsights.com/company/pb-vision ; SwingVision pickleball version: www.sportsbusinessjournal.com/Articles/2023/12/05/swingvisio…
- Our library: appstore.app_day / app_profile (2026-11-19)