Privacy Policy
This policy explains what data our Android apps handle, why, and what you can do about it. It applies to both apps:
- Ai Teacher (package
com.antonchuraev.aiteacher) — an AI teacher for courses on any subject; - Talkstep (package
com.antonchuraev.languagetutor) — an AI tutor for practicing a foreign language.
Both apps work the same way and use the same servers and services. In this policy, "the app" means either of them.
- The app has no account. It does not ask for your name, email address, phone number, contacts or precise location.
- The app shows no ads. To measure our own ad campaigns for the app, Google Analytics for Firebase collects the advertising ID and a few app events.
- Your data is not sold.
- What you write to the AI teacher is sent to our servers and to an AI model so that the teacher can answer.
- In a voice call, the microphone is used only while the call screen is open. The app does not record or store audio and does not send it to our servers; only the recognized text is sent.
- Materials you import are private to you, are never published, and are not used to train AI models.
Who is responsible
Ai Teacher and Talkstep are made and operated by an individual developer, Anton Churaev. For any privacy question or request, write to churaevanton@gmail.com.
How the app identifies you
The first time the app contacts our server, Firebase Authentication automatically creates an anonymous identifier for your installation. It is not tied to your name or email address.
Server records, your subscription status and analytics are linked to this anonymous identifier.
What data is processed and why
1. Lessons and chat with the AI teacher
To run lessons, we process:
- the messages you send to the AI teacher and the teacher's replies;
- your answers to check questions;
- the teacher style, level and learning goal you choose, including the optional free-text goal.
This data is stored on Google Cloud servers (Cloud Functions in the europe-west1 region and Firestore in the eur3 EU multi-region) and is sent to the Google Gemini model through Vertex AI to generate the teacher's reply. The purpose is to make the app work.
1a. Voice calls
In both apps you can talk to the AI teacher by voice. For this, the app asks for permission to use the microphone (Android permission RECORD_AUDIO). You can refuse it and keep using the text chat.
- When the microphone is used. Only during a call, while the call screen is open. If you leave the screen or lock the phone, the call pauses and the microphone is closed. The app does not listen in the background.
- Speech recognition. Your speech is turned into text by the speech recognition built into your device. If on-device recognition is available for the language, it runs on the device. Otherwise the device's system speech recognition service (for example, Google's) is used; that service is provided by the maker of your device or of the service under its own terms, not by us.
- No audio is kept. The app does not record your voice, does not store audio files and does not send audio to our servers or to the AI model.
- What goes to our servers. Only the recognized text of what you said, the language of the call and technical details of the call (for example, whether recognition ran on the device, the recognition confidence, and whether you edited the text before sending). This text is processed like a chat message in section 1: it is stored in the course chat and sent to the AI model to generate the reply.
- The teacher's voice. The teacher's replies come from our servers as text and are read aloud by the text-to-speech engine on your device.
- Interrupting the teacher. While the teacher is speaking, the app can measure how loud the sound from the microphone is, on the device, to notice that you started talking and stop the teacher. This sound is not recorded and is not sent anywhere.
- Call minutes. Voice time is limited. Our server counts the minutes you use against your anonymous identifier and keeps a record of each call: the course, which app it was made from, start and end time, the seconds counted, its status and the call summary.
- Call summary. At the end of a call, the server may send the text of the call to the AI model once to prepare a summary (in Talkstep: suggested corrections and new words). This feedback is generated by AI and is not checked by a person.
You can delete the text of your calls together with the chat history of a course (see "Retention and deletion").
2. Materials you import
You can build a course from your own material: a PDF, photos of pages, pasted text or a link.
- Files are uploaded to Cloud Storage and passed to the AI model to extract the text and build a course.
- The original files are deleted after the course is built, and also if the build fails.
- The extracted text and the built course are stored until you delete the import.
- If you import by link, our server downloads the page you specified.
Imported materials are private to you. They are not published in the catalog and are not used to train AI models.
3. Feedback and reports
- An optional survey after a lesson: the ratings you give on its scales.
- Reports about a teacher's reply: the text of the reply, the reason you pick and your comment.
We use this to improve quality and to moderate AI-generated content.
4. Subscription
Purchases are made through Google Play Billing. Your subscription status and purchase history are processed by RevenueCat, which acts as our processor, and we store the resulting access status. We do not receive your payment card details.
5. Product analytics
We use Amplitude, whose servers are in the United States, to understand how the app is used. Analytics data is linked to the anonymous identifier and includes:
- app usage events: screens, lesson steps, lesson start and completion, paywall views;
- voice call events: call start, pause and end, call length, number of turns, interruptions, reached limits and errors — without the text of what you said and without audio;
- which app is used (Ai Teacher or Talkstep), app version, language, subscription status and lesson counters;
- the course, goal and teacher preset you choose during onboarding;
- the text of a topic request, if you send one;
- technical device data: model, OS version, language and carrier;
- your IP address, from which the service derives an approximate location (country, region, city).
The app currently has no switch to turn analytics off.
5a. Measuring our ad campaigns
We advertise the app with Google Ads. To learn which ads bring people who actually use the app, we use Google Analytics for Firebase. It processes:
- the advertising ID of your device and a Firebase app instance identifier;
- these app events: onboarding completed, lesson started, import submitted, paywall viewed — with details such as the course ID, whether the lesson is free, the type of imported material and what opened the paywall;
- events the SDK records automatically, such as first app open, app sessions, screen views and app updates;
- subscription events (trial start, purchase, renewal), sent from RevenueCat to Firebase;
- technical device data and an approximate location derived from your IP address.
Google may use this data to attribute app installs and these events to our ads. In the European Economic Area, the United Kingdom and Switzerland, the app does not grant Google consent to use this data for ad personalization or to send it for advertising purposes.
You can reset or delete the advertising ID at any time in your Android settings (Settings → Privacy → Ads, or Settings → Google → Ads).
6. Crash reports
Release builds of the app use Firebase Crashlytics. A crash report contains the stack trace, device model, OS version and a Crashlytics installation identifier.
7. Protection against abuse
- Firebase App Check with Play Integrity checks that requests come from a genuine copy of the app.
- We keep request rate-limit counters.
- Server logs (Cloud Logging) contain the anonymous identifier.
8. Time zone and language
- Your device's time zone (IANA identifier) is used to work out when the "next day" starts for a homework task.
- The app language is used for localization.
9. Open textbook search
When you search for open textbooks, your query goes through our server to the Wikimedia API (Wikibooks and Wikiversity) without your user identifier.
Data kept on your device
The app keeps lesson progress, drafts, settings and a chat cache on your device. The state of your lessons and steps, including your answers, is also stored on our servers so that a lesson can continue.
Android backup is turned off for the app. If you uninstall the app, this local data is lost.
Who receives data
- Google — Firebase, Google Analytics for Firebase, Google Ads (measurement of our ad campaigns), Google Cloud, Vertex AI and Google Play.
- RevenueCat — subscription status, purchase history and the Firebase app instance identifier, so that subscription events can be passed to Google Analytics for Firebase.
- Amplitude — product analytics.
- Wikimedia — only the open textbook search query.
Your data is not sold.
International transfers
Data may be transferred to and processed in countries other than the one you live in, including the United States.
Security
Data sent between the app and the servers is encrypted in transit (HTTPS/TLS). Access rules prevent the app on one device from reading another user's data.
Retention and deletion
In the app you can:
- delete chat history, for one course or for all courses;
- delete one import or all imports. This removes the course, the files and the extracted text.
How long data is kept:
- technical idempotency records — 30 days;
- subscription events — 30 days;
- everything else — for as long as it is needed to run the service, or until you ask for it to be deleted.
To delete all data linked to your anonymous identifier, write to churaevanton@gmail.com. The app does not yet have a separate button for this.
Because the app has no account, we may need extra details to find your records, for example the date and time of a purchase or the Google Play order ID.
Uninstalling the app erases the data on your device but does not erase the data on the servers.
Children
The app is intended for users aged 13 and older. It is not directed at children under 13, and we do not knowingly collect data from children under 13.
If you are a parent or guardian and believe that your child has provided data, write to churaevanton@gmail.com.
If you are under the age of majority in your country, use the app with the consent of a parent or guardian.
Your rights
You can ask to access, correct or delete your data, and you can object to its processing. To do so, write to churaevanton@gmail.com.
If you are in the European Economic Area or the United Kingdom, the legal bases for processing are:
- performance of a contract — running the service and the subscription;
- legitimate interests — analytics, measurement of our ad campaigns, crash reports and protection against abuse.
You also have the right to lodge a complaint with your data protection supervisory authority.
Changes to this policy
Changes are published on this page with a new effective date.
Contact
Anton Churaev — churaevanton@gmail.com