Privacy
Last updated: 2026-11-14
Your learning data
Which kanji you have started, your review history, your settings and your pinned example words are stored on your device. The app has no account and never asks for a name, an email address or a phone number. Without a subscription, the app sends none of this data anywhere, but for the strokes you report, described under Stroke reports. If you have turned on your device's own backup (iCloud or Google), the data is part of that backup, under your Apple or Google account.
Sync
Sync is part of the subscription. It keeps your deck the same on all your devices and keeps a copy of it on our server. To do so, the app sends the server your learning progress: when you reviewed each kanji and how it went, which kanji you picked to learn next, your pinned example words, and your settings except those of the device itself, such as the theme. Each request also names the app version and whether it runs on iOS or Android. The connection is encrypted.
The server keeps your data under identifiers that RevenueCat generates for your subscriptions, not under a name. To find it, the app sends the random identifier described under Subscriptions with every sync, and the server asks RevenueCat which subscriptions belong to it. The server keeps only a hash of that identifier, from which the identifier cannot be worked out. It also keeps when a device last synced, updated at most once a day, and the date on which the paid period ends.
The synced data is therefore pseudonymous rather than anonymous: through a subscription identifier it can be connected to the subscription record at RevenueCat. It is stored on the server in readable form, not end-to-end encrypted, because the server has to read it to combine what your devices send and to optimize the review schedule.
When the subscription ends, the data stays on the server and the devices that synced it can still fetch it. The server then accepts no new data. On Android, a device set up as new after the subscription has ended cannot reach it while none of the devices that synced before is left, even with a new subscription, as Google Play shows a device only the subscriptions that run. The legal basis is the subscription contract (Art. 6(1)(b) GDPR).
Optimizing the review schedule
The scheduler computes when a kanji comes up for review from a set of numbers, its parameters. The app ships with default parameters. While your subscription runs, our server optimizes the parameters for your review history and syncs them to your devices, which then schedule with them. This is part of the subscription (Art. 6(1)(b) GDPR).
During that time the server also uses your review history, together with those of other subscribers, to check how well this works and to compute better default parameters for the app. Defaults are computed from at least 20 decks, and only the optimized parameters leave the server, never a review history. The legal basis is our legitimate interest in improving the schedule for everyone (Art. 6(1)(f) GDPR).
You can object to this second use at any time. Turn off "Help optimize the default schedule", under Privacy in the app's settings, where it is shown once you have subscribed. From your next sync on, your review history is left out of it.
Stroke reports
When the app refuses a stroke you drew right, you can report it from the card. Once you confirm, the app sends our server your strokes of that kanji, which kanji it is, why the stroke was refused, the strictness you have set, whether strokes are corrected as you draw, the app version, whether it runs on iOS or Android, and the version of the stroke data it judged them against. This works with or without a subscription. A report carries no identifier of you, your device or your subscription, and the server keeps nothing that connects it to you: only the report as it came and when it came. The app deletes a report once the server has taken it.
We fetch the reports to our own computers to test how the app judges strokes, and they are then deleted from the server, which keeps at most 10,000 until then. Deleted reports stay in the server's backups for up to a week. We keep the fetched reports for as long as we use them to improve the stroke check. Since a report cannot be connected to you, we cannot find or delete one for you. The legal basis is our legitimate interest in judging strokes correctly (Art. 6(1)(f) GDPR).
Error reports
When something goes wrong in the app, it sends our server an error report: the kind of error, the first line of its message, the places in the app's code where it happened, the app version, whether it runs on iOS or Android, the version of the operating system, and the version of the kanji data built into the app. Before the app keeps a report, it takes out file paths, identifiers and any Japanese text, so a report carries no identifier of you, your device or your subscription, and none of your kanji or words. The server keeps nothing that connects it to you: only the report as it came and when it came. The app reports the same error at most once a day, keeps at most 50 reports until they are sent, and deletes a report once the server has taken it. Where the app cannot start, it sends the report at once and keeps nothing. This works with or without a subscription.
We fetch the reports to our own computers to find and fix faults in the app, and they are then deleted from the server, which keeps at most 10,000 until then. Deleted reports stay in the server's backups for up to a week. We keep the fetched reports for as long as we need them to fix the faults they show. Since a report cannot be connected to you, we cannot find or delete one for you. The legal basis is our legitimate interest in an app that works without faults (Art. 6(1)(f) GDPR).
You can object to this at any time. Turn off "Send error reports", under Privacy in the app's settings. From then on the app neither keeps nor sends error reports, and it deletes those it has not sent yet.
Subscriptions
The subscription is sold and billed by Apple on iOS and by Google on Android. Your payment details go to them and never to us. Both process that data under their own privacy policies.
The app uses RevenueCat to find out whether a subscription is active. It sends RevenueCat a random identifier created on your device. When you sync, our server sends RevenueCat the same identifier and receives the subscriptions that belong to it. If you subscribe, RevenueCat also receives what the store reports about the purchase: which product was bought, when it renews and whether it is still active.
Each request from the app also carries the app version, device model, operating system version, preferred languages and store country, and, on iOS, the identifier that iOS gives your device for our apps (the identifier for vendor). RevenueCat sees the IP address the request comes from. The app also reports when the subscription screen is opened. RevenueCat receives no store account and none of your learning data. Its own policy is at revenuecat.com/privacy. The legal basis is the subscription contract (Art. 6(1)(b) GDPR) and, before a purchase, our legitimate interest in knowing whether one exists (Art. 6(1)(f) GDPR).
Analytics and advertising
The app contains no advertising, and it does not track you across apps or websites. Apart from the learning data, the stroke reports, the error reports and the opening of the subscription screen described above, it does not record what you do in it. If you have agreed in your device settings to share diagnostics with app developers, Apple or Google sends us crash reports. They contain no learning data and do not identify you.
This website
These pages set no cookies, use no analytics and load nothing from other domains. The hosting provider, Cloudflare, processes what every request carries (IP address, time, requested address, browser identification) to deliver the page. The legal basis is our legitimate interest in operating the website (Art. 6(1)(f) GDPR).
If you write to us, we use your email address and your message to answer you, and delete both once the matter is closed. The legal basis is our legitimate interest in answering requests (Art. 6(1)(f) GDPR).
Processors and transfers outside the EU
Three companies process data on our behalf, each under a data processing agreement that obliges it to keep the data confidential and to use it only to provide its service.
- Amazon Web Services EMEA SARL, Luxembourg (servers in the EU)
- Cloudflare, Inc., United States
- RevenueCat, Inc., United States
Cloudflare is certified under the EU-US Data Privacy Framework. Transfers to RevenueCat take place under the European Commission's standard contractual clauses.
Storage period
The website and the sync server keep no logs of requests. The sync server counts its requests per day by their outcome, with nothing that tells one deck or sender from another. To limit how often one sender can ask, the sync server holds the address a request comes from in memory for a short time. The synced data is kept until you turn off sync in the app's settings, which deletes it. It is also deleted automatically after 24 months in which none of your devices has synced. Subscription records at RevenueCat are kept for as long as the subscription and the statutory retention periods for accounting records require.
The server is backed up, on its own disk and as a copy at the same provider in the EU. Deleted synced data stays in these backups for up to a week. Should the server ever be restored from one, synced data deleted shortly before would be on the server again: turning sync on and then off again deletes it.
Deleting your data
Deleting the app deletes the learning data on that device. "Turn off", under Sync in the app's settings, deletes everything our server holds for your subscription: the synced data, the subscription identifiers and the hashes of your devices' identifiers. Sync is then off on all your devices, and each keeps its learning data. It stays off on a device until you turn it on there again, and a newly installed app syncs again. Turning off also works after the subscription has ended. We do not delete synced data in response to an email: a Support ID or an order number does not show us that the sender is the subscriber.
The subscription record at RevenueCat is the one thing we delete for you. Turn off sync first: the server finds your data through that record, and once the record is gone the app can no longer reach the synced data, which is then deleted only as described under Storage period. To have the record deleted, send the Support ID from the app's settings, or the order number from the store's receipt, to kaisho@gruber.software. Two things are kept: the record of a subscription that is still running, until it ends, and what the statutory retention periods for accounting records require, until they end.
Your rights
Under the GDPR you have the right to access, rectification, erasure, restriction of processing, data portability and objection. You object to the use of your review history for the app's defaults, and to the error reports, in the app itself, as described under Optimizing the review schedule and Error reports. If something is wrong, please contact us first. You can also complain to the Austrian data protection authority, the Datenschutzbehörde, or to the authority of the country you live in.
The only data we can look up for you is the subscription record at RevenueCat, and for that we need the Support ID or the order number. The synced data is the same data as on your device, and "Save a backup" in the app's settings gives it to you as a file. We do not send it by email, for the reason given under Deleting your data.
Contact
Kaisho and this website are run by Clemens Gruber. The address is in the imprint. For questions about privacy, write to kaisho@gruber.software.
