See what keeps happening in the points you lose.
Import a video of your own match. Patemo finds the rallies, you check them and note what happened, and it shows the situations where you lose most, with every clip behind the number.
Free while Patemo is in development. Nothing is for sale yet. No account, no upload: your videos stay on your device.
Find the rallies
Import a match from Photos or Files. Patemo listens for the ball and proposes where each rally starts and ends.
Note what happened
Won or lost, whose serve, where the point was decided. Each label takes a tap, and any of them can stay open.
See what keeps happening
Patemo counts your own labels by situation and shows every clip behind each number.
Rallies found, and yours to correct
Patemo listens for the ball on the table and the bat, on your device, and marks each rally it hears. Trim a start or end, split two rallies, join two, add one it missed or delete one. Every change can be undone, and finding rallies again never overwrites your corrections.
Slow motion, frame by frame and a player that stops at the end of the rally make it quick to see what really happened.
A few taps say what decided the point
Mark whether you won or lost, who served and how long, the phase it was decided in (serve, receive, third ball or rally), how it ended and which side your last stroke came from. Leave anything you are unsure of empty: Patemo never fills it in for you.
A number you can check, clip by clip
Pick a situation, such as receiving a short serve. Patemo shows how many comparable points you lost, the range the real share is likely in, and every clip: the points you lost and the ones you won. It also says which points were left out and why.
With fewer than eight comparable points it says there is not enough yet and draws no conclusion. It shows a pattern in your points, not its cause.
Practise it, then measure again
Each situation links to drills written for it, with a table diagram for every step. There are 16 today, for training with a partner, a robot or on your own.
Set a goal and Patemo keeps the numbers you start from. Only matches you played after that, in the same format, are compared with them, on the same counting rules. Matches that do not count are listed with the reason.
Bigger screen, same match
On an iPad or Android tablet the video and the rally list sit side by side. Patemo looks and works the same on iPhone and Android.
Stays on your device
Patemo has no account and sends nothing anywhere. Your videos, corrections and labels stay in the app on your device. The original in Photos or Files is only read, never changed. Delete one match or everything whenever you want.
Export the rallies you choose as a new video, without a watermark, and share them yourself.
Questions
What kind of video works best?
A whole match filmed from a phone that stays still, from behind or beside the table, with the sound on. Patemo listens for the ball, so a quiet recording or a busy hall makes it miss more. The support page has a short filming checklist.
Does Patemo judge my technique?
No. It does not recognise strokes, movement or where the ball lands. The situations come from the labels you set, and Patemo counts them. It shows where points go, not why.
Is my video uploaded?
No. There is no account and the app makes no network requests. Everything stays on your device until you share an exported clip yourself.
Does it work for doubles?
You can mark each match as singles or doubles. A goal only compares like with like, so a doubles match never counts against a singles starting point.
What does it cost?
Nothing yet. Patemo is still in development and not yet in the App Store or Google Play. Nothing is for sale.
Your own matches, point by point.
How Patemo is made
We develop Patemo with AI tools for programming and translation. We use automated checks and device tests. The Patemo team is responsible for the result, and a person answers support requests.
AI and data processing in this app
Patemo finds rallies by listening for the ball in your video, on your device. To suggest forehand or backhand it runs an on-device body-pose model (Apple Vision on iPhone and iPad, Google's open MoveNet on Android). Nothing is uploaded, and every result can be corrected.