What English tutor does with your information — in plain terms and in full.

01Overview

English tutor helps you learn English from whatever language you already speak — Arabic, German, Spanish, Chinese or any other: you write a sentence and it is corrected and explained in your language, you photograph a page and every word on it becomes tappable, you record yourself and it is transcribed, and the words you meet collect into a list you can revise.

Most of that happens on the device itself. The parts that need a large language model — grammar reasoning, translation nuance, exercise reading, progress review — go to Google's Gemini API using an API key that you supply. Everything in this policy is about that boundary: what stays, what goes, and what you can turn off.

English tutor has no user accounts, no back end and no servers operated by its developer. The developer never receives your sentences, your recordings, your photographs or your word list.

02What stays on your device

All of the following is written to local app storage and never leaves the device unless you switch on Drive sync (§09):

• Your saved words — the headword, meanings, grammatical forms, your own notes and example sentences, and the spaced-repetition schedule for each one.
• Your practice history — the sentences you checked, what was corrected and when, which is what the progress statistics and the "your topics" list are built from.
• Books and study plans — any PDF you import stays in the app's own storage, together with extracted page text, page images and the day-by-day plan.
• Photographs you choose — a picture you pick or take for a page or an exercise is held for as long as that view needs it.
• Recordings — audio you record is written to a local file so it can be played back and transcribed.
• Settings — including your Gemini API key and, if you use it, your Google OAuth client details.
• A response cache — so that checking or translating the same text twice does not cost a second request.

The bundled English dictionary ships inside the app and is read locally. Looking a word up sends nothing anywhere.

03Your own Gemini API key

The AI features are not a service the developer runs on your behalf. You create your own Google Gemini API key and paste it into Settings, and the app then calls Google's API directly from your device as you.

Two consequences worth being plain about:

• The key is stored in the app's own settings storage on your device only. It is not transmitted to the developer and there is nowhere for it to be transmitted to.
• Your use of the Gemini API is governed by your own agreement with Google. How Google handles what is sent is described in Google's terms for the Gemini API and in the Google Privacy Policy. Please read them — free-tier API usage in particular may be used by Google to improve its products.

If you never add a key, every AI-backed feature simply reports that it needs one, and the offline half of the app — the dictionary, tap-a-word lookup, word forms, your saved list, the review cards, the PDF reader, on-device OCR and on-device transcription — keeps working.

04What is sent to Gemini

Only when you press the button for the feature in question, and only the material that feature needs:

• Checking a sentence — the English text you typed, and the name of your own language (for the explanation).
• Translating — the text you typed, in English or in your own language.
• Asking the tutor — your question and the recent conversation.
• Reading an exercise or a page — wherever the device could read the text itself, the extracted text is sent rather than the picture. Some picture-based features do send the image itself, and those are listed in §05.
• Transcribing a recording — see §06: the audio is sent only when on-device transcription is switched off or unavailable.
• A progress review — summary statistics from your practice history and a sample of your saved words and recent corrections.

Requests go over HTTPS to generativelanguage.googleapis.com. The app keeps a local cache of responses so repeating the same request does not send it again.

05Camera, photos and page reading

Permission: Camera

The camera is used in two places, and only when you start them: taking a picture of a page or an exercise, and scanning the pairing QR code shown by the desktop app (§10). English tutor does not open the camera at any other time.

A picture you take or choose is processed like this:

• On-device OCR first. Text recognition runs locally on Android through Google's ML Kit text recognition, which is bundled with the app and does not send the image anywhere. That is what makes every word on a photographed page tappable.
• Text, not pictures, where possible. For reading the tasks on a page, the locally extracted text is what gets sent to Gemini.
• Pictures where the picture is the point. Some features do send the image itself to Gemini — reading an exercise whose layout matters, and answering questions about a photographed page. You are choosing that feature explicitly when it happens.

Photographs are not uploaded anywhere else, are not added to any gallery by the app, and are not retained after you clear or replace them in the view.

06Microphone and speech

Permission: Microphone

The microphone is used only while you are recording in the Record or Live views, after you press record. It is never listening otherwise.

• On-device transcription is the default path. On Android, speech recognition runs locally with a Whisper model through sherpa-onnx; on Windows it runs locally through faster-whisper. In this mode your audio never leaves the device — at most the resulting English translation of the text is sent on.
• Sending audio to Gemini is the fallback. If local transcription is switched off or is not available, the recording is sent to the Gemini API to be transcribed. This uses your own API key, as in §03.

Recordings are saved as files in the app's own storage so you can play them back, and you can delete them from within the app.

07Reading English aloud

You choose in Settings › Pronunciation which voice reads English aloud. Each one is used only when you press a speaker button, and each receives only the word or sentence being read:

• System — your device's own text-to-speech engine. Whether that engine works offline is a property of the engine you have installed.
• Piper — an open-source voice that runs entirely on your device. Nothing is sent.
• Google — the text is sent to Google's public text-to-speech endpoint (translate.google.com) and the audio comes back.
• Azure — only if you add your own Microsoft Azure Speech key. The text goes to Microsoft's speech service in the region you chose, under your own Azure agreement and the Microsoft Privacy Statement. The key stays on your device.

08Signing in with Google

Google sign-in is optional and exists for one purpose: syncing your word list through your own Google Drive. The app works fully without it, and nothing is gated behind it.

When you do sign in, the app receives an access token for your account and the scope described below. Your Google password is never seen by the app — the sign-in happens through Google's own flow.

09Google Drive word-list sync

Scope: drive.file only

English tutor asks for https://www.googleapis.com/auth/drive.file and nothing broader. That scope grants access only to files the app itself creates. The rest of your Drive is invisible to it — the app cannot list, read, or modify documents it did not make.

What is synced is your vocabulary list, so that the phone and the desktop app show the same words. It is stored in your Drive, under your own Google account. The developer has no access to it.

The narrow scope is deliberate: broader Drive scopes are classed by Google as sensitive or restricted, and this app has no need of them.

10Pairing with the desktop app

If you use English tutor on a computer as well, the two can be paired by showing a QR code on one and scanning it with the other. This exists so you do not have to retype an OAuth client ID and secret and a Gemini API key on the second device.

What moves across is your own credentials and settings, between two devices you own. It does not pass through any server belonging to the developer. Treat the QR code as you would a password: anyone who photographs it while it is on screen gets what it carries.

11Downloaded models, and the full version

The on-device Android speech model and the Piper voices are too large to ship inside the app, so the first time you use them they are downloaded once from the sherpa-onnx project's public release files on GitHub and cached on your device. That request fetches a file; it carries none of your content. Ordinary network metadata such as your IP address is visible to GitHub as it would be for any download, under GitHub's privacy statement.

On Android the app also reads one small public settings file from maiaarchitech.com (whether the daily free allowance applies). The request carries nothing about you.

Purchases — the one-time full version and any optional support — are handled entirely by Google Play billing. The app learns only that a purchase exists; payment details never reach the app or the developers. On Windows the app is bought on the Microsoft Store, under Microsoft's policies.

12Security

• Every network request the app makes is over HTTPS.
• Your data sits in the app's private storage area, protected by the operating system's own app sandbox and by your device's lock screen and encryption.
• Because there is no server and no account, there is no central store of user data to be breached.
• Your Gemini API key and any OAuth secret are held in local app storage. Anyone with unlocked access to your device could reach them, which is the ordinary reason to keep a device locked.

13Keeping and deleting your data

English tutor keeps your data for as long as the app is installed, because the data is the app — there is no separate copy held anywhere else.

• Delete individual items in the app — words can be deleted from the list, recordings and photographs cleared from their views, books removed.
• Delete everything — uninstalling the app, or clearing its storage in Android Settings, removes the local database, cached responses, downloaded models, your API key and your saved settings.
• Revoke Google access — at myaccount.google.com/permissions, which withdraws the app's Drive access.
• Remove synced files — anything the app put in your Drive can be deleted by you from Drive.
• Data already sent to Google — requests made with your Gemini API key are subject to Google's own retention, and are dealt with through your Google account, not through this app.

There is no account to close and no request to make of the developer, because the developer holds nothing of yours.

14What English tutor does not do

15Third parties, at a glance

16Children's privacy

English tutor is a general-audience study tool aimed at adult and teenage learners of English. It is not directed at children under 13, and it does not knowingly collect anything from them — it does not knowingly collect personal information from anyone, since it has no server to collect it to. Setting up the AI features requires obtaining a Google API key, which is an adult step.

17Changes to this policy

If the app's handling of data changes, this page is updated and the effective date at the top changes with it. Material changes will also be noted in the app's release notes on Google Play. The current version of this page always describes the current released version of the app.

18Contact

Questions about this policy, or about anything the app does with your information:

Ahmed Yosri
ahmed.yosri.u3@gmail.com