Tono Translate is a translation and language-learning app. This policy explains what information the app handles and where it goes. The short version: your translations, lessons and feedback requests pass through our servers to an AI provider and are not stored there; your library lives on your device; there are no ads in the app, we don't track you across other companies' apps and websites, and we never sell your data. Outside the EEA, the UK and Switzerland, the app tells Meta when it is opened, so we can see which of our ads on Instagram and Facebook bring people to it.
What we collect
Account. Tono Translate does not ask you to register. On first launch the app creates an anonymous identifier for your device so your free allowance and your subscription can be recognized; it contains no name, email or Apple ID. It is shared only with the companies that run the app for us, listed under Who receives it, below, and RevenueCat uses it as your subscriber number. Your subscription itself belongs to your Apple ID, not to us.
Proof of a genuine app. To keep the free allowance from being abused, the app asks Apple's App Attest service to confirm that it is a genuine copy of Tono Translate on a real device. We keep the key identifier Apple returns with your anonymous identifier. It identifies the app installation, not you.
Content you submit for translation, feedback or lessons. Phrases you type, context you provide, photos you take of text, and transcriptions of speech are sent to our servers and forwarded to our AI provider (currently Google's Gemini API) to produce the translation. When you turn on live listening in Talk, the sound from your microphone is sent the same way, so our AI provider can turn it into text in either language as it is spoken. When you ask for feedback on a sentence you wrote, that sentence is sent the same way, and so is the material a weekly lesson is written from (listed under Your library, below). We do not store this content on our servers — it passes through to produce your answer, and nothing we keep contains what you translated or submitted.
Our AI provider processes it under its own terms, including Google's API data-use policy. We use Google's paid API tier, under which submitted content is not used to train or improve Google's models. If we ever change providers or tiers in a way that affects this, we will update this policy.
Phrases Tono Translate writes for you to learn. When the app writes a phrasebook or new practice phrases for you, our AI provider writes them from the situation you describe or the phrases you looked up, and a second AI provider, Anthropic's Claude API, checks the written phrases for natural wording before they reach you. Anthropic receives only those written phrases, with what each one means and who it is said to, never what you typed, looked up or translated. Under Anthropic's commercial API terms, content sent to it is not used to train its models. We do not keep any of it on our servers; the phrases are kept on your device.
Where processing happens. Our servers and our AI providers operate in the United States and other countries, so what you submit for translation, feedback or lessons, and the phrases you listen to, are processed outside your own country, including outside the EU/EEA and the UK.
Speech. Speech is converted to text by Apple's speech recognition, on your device where supported. On some devices Apple may process the audio on its servers under Apple's privacy policy. Only the resulting text is sent for translation, and only when you ask for one. The exception is live listening in Talk, which you turn on yourself: while it is on, the sound from your microphone passes through our servers to our AI provider to be turned into text, and we do not record or store it. It stops when you tap Stop or leave Talk, and the microphone pauses while Tono Translate is speaking.
Listening to a phrase. When you tap Listen in the app, or show a phrase in Show Mode, and you are online, the phrase's text is sent through our servers to Google's Cloud Text-to-Speech service, which returns it as audio in a natural voice. We do not store the text or the audio on our servers. The audio is kept on your device, so the phrase plays again without being sent a second time, and it is deleted when you erase your library. In Talk and face-to-face, and whenever you are offline or have turned on Translate on this device, Tono Translate uses the voice built into your device instead, and nothing is sent.
Your library. Translation history, scripts, and starred phrases are stored locally on your device, not on our servers. Weekly lessons, practice cards and your practice history, and the phrases shown in the Home Screen widget are also stored only on your device. So is the audio of phrases you've listened to. A weekly lesson is generated the same way as a translation: the week's phrases and script titles, last week's lesson summary, and the goal and level you set are sent through our servers to our AI provider to write the lesson, and are not stored there. If you turn on conversation saving in Talk (off by default), full transcripts — including the other person's words — are also kept on your device only, with a visible indicator while saving is on, until you delete them.
Usage counts. We record how many translations are made per day against your device's anonymous identifier — this enforces the free allowance and helps us understand load. For live listening in Talk, we also record how many seconds you use each month, to apply the monthly allowance. For phrases read aloud, we record how many characters are read each day, to keep within a daily limit. Counts only; not the content. We also count, per day, how many people reach each step of the app's first-time setup, how many open the app again a day and a week later, and how many see each subscription offer — a tally of numbers with no identifier attached, so it cannot be connected to you.
Subscriptions. Purchases are processed by Apple; we never see your payment details. We use RevenueCat to manage subscription status, which processes your Apple transaction identifiers under its own privacy policy. When you tap to subscribe, the app also notes which screen you were on (for example, the offer at the end of first-time setup, or the Practice tab), and RevenueCat keeps that note with your subscription, under your device's anonymous identifier. We use it only to see which parts of the app lead people to subscribe.
Server logs. Like any web service, our server host keeps standard request logs, including IP addresses, for a few days, for security and to prevent abuse (for example, limiting how many new identifiers one network can create in an hour).
Measuring our ads. We advertise Tono Translate on Instagram and Facebook. To see which of those ads bring people to the app, the app includes Meta's measurement code, which tells Meta each time you open or return to the app, along with general details about your device. It never reads your device's advertising identifier, and it sends nothing you translate, type or keep in the app. It doesn't run at all if your App Store account is in the EEA, the UK or Switzerland. Exactly what Meta receives is listed under Who receives it, below.
Who receives it, and where
A few companies run parts of Tono Translate for us, and one measures our ads. Each receives only what its job needs, over an encrypted connection:
- Google LLC (United States), the Gemini API: the text, context and photos you submit, and the sound from your microphone while live listening is on, to write your answer; and Cloud Text-to-Speech: the text of a phrase you listen to, to read it aloud. Not kept by us, and not used to train Google's models.
- Anthropic, PBC (United States), the Claude API: the phrases the app writes for you to learn, with what each one means and who it is said to, to check that they sound natural. Not what you typed or translated, not kept by us, and not used to train Anthropic's models.
- Supabase, Inc. (United States), our server host: your anonymous identifier, your counts and subscription status, the App Attest key identifier and request logs. What you submit passes through it on its way to Google.
- RevenueCat, Inc. (United States), subscription management: your anonymous identifier, your Apple transaction identifiers, and the screen you subscribed from.
- Apple, for purchases, App Attest, and speech recognition when it isn't done on your device, under Apple's privacy policy.
- Meta Platforms, Inc. (United States), to measure our ads on Instagram and Facebook: each time you open or return to the app, general details about the device and the app (such as its model, iOS version, language, time zone, screen size, storage space and the app's version), an identifier Meta's code creates for this app on your device, and your IP address, as with any connection; and, if Meta's code itself fails, a report of the error. Apple also tells Meta, without identifying you, when an install came from one of our ads (Apple's SKAdNetwork). Meta doesn't receive your device's advertising identifier, which the app never reads, or anything you translate, type or keep in the app. Nothing is sent if your App Store account is in the EEA, the UK or Switzerland. Meta handles what it receives under its own terms and privacy policy.
To keep what you translate on your iPhone, turn on Settings → Translate on this device: Translate, Talk, Show Mode and face-to-face then translate on the device, with the words only and none of the explanations, live listening is off, and Listen uses your iPhone's own voice. Photos, Polish, Ask about this, Scripts, weekly lessons, the share sheet and Siri still use our servers.
Why we use it
We use what you submit to give you the answer you asked for, and your identifier, counts and subscription status to provide the service you chose (in EU and UK terms, to perform our contract with you). We use the App Attest check, the request logs, the setup tally and the screen you subscribed from in our legitimate interest in preventing abuse, keeping the service running and understanding what works. Outside the EEA, the UK and Switzerland, we measure our ads with Meta in our legitimate interest in knowing which ads bring people to the app. The website's analytics cookie in the EEA, the UK and Switzerland, and the app's notifications, depend on your consent, which you can withdraw at any time.
How long we keep it
What you submit is not kept: it exists on our servers only while your answer is being written. Your anonymous identifier, with its counts, subscription status and App Attest key identifier, is kept while you use the app. When it has gone 12 months without use or a subscription, we delete it automatically, and you can ask us to delete it sooner. Request logs are kept for a few days. The setup tally is kept as totals only. RevenueCat and Apple keep subscription records under their own policies. Meta keeps what it receives under its own policy. It isn't connected to the identifier we keep, so we can't look it up or delete it for you; Meta's privacy policy explains how to ask Meta.
Website analytics
Our marketing website (tonotranslate.com, including this page) uses Google Analytics to count visits and see which pages people find useful. It records page views, the site you came from, approximate location derived from a truncated IP address, and general device and browser type, and it sets a cookie in your browser. This is about website visitors, not app users: it is not connected to your Tono Translate account. The app's own measurement of our ads is described under Measuring our ads, above.
If you're in the EEA, the UK or Switzerland, that cookie is not set unless you accept it — we ask before anything is stored, and declining is one click. You can change your answer at any time with Cookie settings, at the bottom of every page. We do not use advertising cookies anywhere.
What we don't do
- No ads in the app, and the app never reads your device's advertising identifier.
- No tracking of you across other companies' apps or websites. The app never shows Apple's request to track you, so Meta's code tells Meta that tracking isn't allowed.
- No sale of any data, ever.
A note on other people
Tono Translate's conversation features can transcribe what someone speaking another language says to you. That transcription is processed the same way as anything else you translate — Apple's speech recognition as described above, or, with live listening on, their voice passing through our servers to our AI provider to be turned into text; text sent for translation; nothing stored by us. If you use these features in a conversation, it's your responsibility to tell the other person, and to get their consent where the law requires it. Please be considerate about using them where recording assistance is inappropriate.
Deleting your data
Your library — every phrase, script and transcript — exists only on your device. Deleting the app deletes it. The widget keeps a copy of up to 30 phrases on your device. After you erase your library, it's cleared the next time you open the app. If you delete the app, iOS may keep this copy until the device is erased. Inside the app, Settings → Erase my library removes the same content without uninstalling, and Reset app returns settings to first launch. If you back up your iPhone to iCloud or a computer, your library is included in that backup. We have no access to it.
What stays on our servers is your anonymous identifier with your translation, live-listening and read-aloud counts, subscription status and App Attest key identifier. It carries no name, no email address and none of what you translated. To have it deleted, use Settings → Delete my server data, which emails us that identifier, or write to hello@tonotranslate.com and we'll help you find it. We delete it within a month. Afterwards the app starts with a new identifier; if you subscribe, restore your subscription from the subscription screen. Canceling a subscription is separate and handled through your Apple account settings.
Your rights
Depending on where you live, including the EU, the UK, Brazil and Korea, you have the right to ask what we hold about you, and to have it corrected, deleted or restricted, to object to how we use it, to receive a copy of it, and to withdraw your consent. Send your request to hello@tonotranslate.com with your device's identifier (Settings → Delete my server data includes it), and we'll answer within a month. You can also complain to your data protection authority. In Korea, that is the Personal Information Protection Commission (pipc.go.kr) or the Personal Information Infringement Report Center (privacy.kisa.or.kr, phone 118).
Children
Tono Translate is not directed at children under 13 (under 14 in Korea), and we don't knowingly collect information from them. If you believe a child has used the app, write to us and we'll delete what our server holds for that device.
Changes and contact
Tono Translate is operated by Howard Daein Kim, who is responsible for this policy and for the personal information it describes. If this policy changes materially we'll note it in the app. Questions and requests: hello@tonotranslate.com.