Lecta Privacy Policy
This policy explains how Lecta ("we", "us") handles personal information when you use the Lecta app (listed in app stores as "Lecta 노트") (the "app"). The Korean version of this policy applies to users in Korea and prevails for them. For everyone else, if the Korean and English versions differ, the English version prevails.
Your recordings and transcripts are stored only on your device by default. We do not keep them on our servers.
1. What we collect and how
| Type | Items | When | Where it is stored |
|---|---|---|---|
| Account | Google or Apple account email (with Apple this may be a hidden relay address), name, account ID (UID) | When you sign in with Google or Apple | Our servers (Firebase) |
| Device | A one-way hash of a device identifier (Android: Android ID; iOS: the app's identifier for your device), app integrity check results, push notification token | When you use cloud features; when you sign in (notification token) | Our servers |
| Usage | Token balance, plan, minutes used and tokens spent per feature, invite code and referral rewards, instant caption connection records, terms consent records (version, items, time, and your age range with the time it was checked), request times | When you use cloud features | Our servers |
| Age check | Your age range (under the sign-up age / sign-up age to 17 / 18 or older) and when it was checked. We do not store your birth year (section 7) | When you first open the app, and when you check again in Settings | Your device. Once you sign in, also our servers, together with your terms consent record |
| Access logs | IP address, access time, error logs | Automatically, when the app contacts our server | Our servers (cloud logs) |
| App crash reports | Error details (where in the code it happened), device model, OS version, app version, a random identifier created for each app installation, and the time. No recordings, transcripts, note content, names or email addresses are sent | Automatically, when the Android app crashes or hits an error | Google (Firebase Crashlytics) |
| Student verification (optional) | University students: your school email address. We use it only to send a verification link; our server stores just a one-way hash of the address and the school domain (for example snu.ac.kr). If the Google account you signed in with is a school address, we verify you with that account directly. Middle and high school students: a photo of your student ID (including the name, student number, face and anything else shown on it), the school name you enter, your account email and the time of the request. We delete the student ID photo as soon as we have reviewed it. If approved: school name, verification method, approval time and expiry date | When you apply for student pricing on our website | Our servers (Firebase) |
| Payments (once web payments launch) | Order number, the payment provider's payment or subscription ID, product, amount, currency, payment status and dates. We never receive card numbers or other payment details; the payment provider handles them | When you pay on our website | Our servers |
| Notifications | Personal notices (tokens added, payment and subscription status, student verification results) and whether you have read them. Whether you agreed to receive offers and events notifications, and the date you agreed or declined | When one of those things happens; when you turn notifications on or off in Settings | Personal notices: our servers. Your consent choice and its date are stored on your device only |
| Lecture materials (optional) | Text your device reads from the lecture materials you choose (PDF, slide or document files). For scanned PDFs with no text, reduced page images. The files themselves are not sent | When you use lecture materials for AI notes | Our server deletes them as soon as the term list is extracted. The extracted term list is stored on your device only |
| Audio | Recording files, live recording audio | When you record or import a file | Your device only |
| Text | Transcripts, corrected transcripts, AI notes, translations, term hints, highlights and bookmarks | When you use the app | Your device only |
Information sent outside your device only when you turn on optional features (details in section 4):
- Cloud transcription: recording audio and term hints are sent to a speech recognition provider. When it is off, speech recognition runs only on your device.
- AI correction, AI notes and translation: depending on the feature, the transcript text, term hints, the field you chose and any note template you created are sent to an AI provider.
- Instant captions: the audio being recorded is streamed in real time to an AI provider. Instant captions are only for people 18 or older (section 7).
- Drive backup: recording files, transcripts, corrected transcripts and AI notes are saved to your own Google Drive. The app can see only the files it created, not your other Drive files.
Speaker separation: telling speakers apart in a recording happens on your device. The voice features used for this never leave your device and are not used to identify anyone.
Sensitive and health information: recordings can contain other people's voices and, for example in medical lectures, health information. We do not intentionally collect health information. Do not use Lecta to record patient care or any conversation containing patient information. Get any consent you need before recording, and do not use cloud features for recordings that contain identifiable patient information.
The app does not collect advertising IDs, location or contacts.
2. Why we use it
| Purpose | Information used |
|---|---|
| Confirming who you are and keeping you signed in | Account information |
| Providing cloud transcription, AI correction, AI notes, translation and instant captions | Audio, transcript text, term hints |
| Extracting specialist terms from your lecture materials so AI notes spell them correctly | Text or page images from the lecture materials you choose |
| Giving and deducting tokens, managing plans | Usage records |
| Sending announcements (maintenance, updates and so on) and notices about finished jobs, tokens, payments and student verification results | Push notification token, personal notices, usage records |
| Sending offers, events and other promotional messages to users who agreed to receive them | Push notification token |
| Limiting features by age (sign-up age; 18 or older for instant captions) | Age range and when it was checked |
| Preventing duplicate free tokens and referral rewards | Device identifier hash, usage records |
| Checking eligibility for student pricing and preventing one school email from being used twice | Student verification request (school email and its hash, student ID photo, school name, account email) |
| Taking payments, giving refunds and keeping transaction records required by law | Payment records, account email |
| Checking the app is genuine and preventing abuse | App integrity results, access logs |
| Fixing errors and answering questions | Access logs, app crash reports, messages you send us |
We do not use your audio or transcripts for AI training, advertising or marketing, and we do not keep them on our servers after processing. See section 4 for how long outside providers keep it.
Promotional messages
We send promotional messages such as discounts and events as push notifications only if you turn on “Offers & events” in the app’s Settings, which counts as your consent. It is off by default.
- Promotional notifications start with “(Ad)” and say how to stop them.
- We do not send them between 9 PM and 8 AM (Korea time).
- When you agree or decline, the app shows you the date, and we ask you to confirm your consent again every 2 years.
- You can stop them at any time in app Settings > Display & general > Notifications > Offers & events.
- Announcements and personal notices (tokens added, payments, student verification results and so on) are not promotional, and you can turn each kind off in the same menu.
3. How long we keep it
| Information | Retention |
|---|---|
| Account information, token balance, terms consent records (including your age range) | Until you delete your account |
| Age range and check time stored on your device | Until you delete the app (checking again in Settings replaces it) |
| Usage records, referral records | Deleted immediately when you delete your account |
| Push notification token, instant caption connection records | Deleted immediately when you delete your account |
| Personal notices | Deleted automatically 180 days after they are created, or immediately if you delete your account before then |
| Device identifier hash | 1 year after last use (to prevent repeat free tokens and referral rewards). If you delete your account, it is unlinked from the account on our server, kept for this period, then deleted automatically. |
| Access logs (IP address, etc.) | 30 days (default server log retention) |
| App crash reports | Deleted automatically after 90 days (Firebase Crashlytics retention). They are not linked to your account, so they are deleted on this schedule whether or not you delete your account |
| Text and page images sent from lecture materials | Deleted from our server as soon as the term list is extracted (our server does not store them; it only uses them while processing). The extracted term list is stored on your device only (until you delete it or uninstall the app) |
| Audio and text sent for cloud processing | Our server discards it as soon as processing is done. Providers keep it as described in section 4. |
| Recordings and transcripts on your device | Until you delete them or uninstall the app (we have no access) |
| Drive backup | Until you delete it from your Drive |
| Student ID photo | Deleted as soon as we review it (approve or reject). A request nobody reviews is deleted automatically, photo included, after 7 days (actual removal can take up to about a day longer) |
| Reviewed student verification requests (result, rejection reason, reviewing admin; no photo) | 30 days |
| Student verification result (school name, school domain, method, approval time, expiry date) and school email hash | Until it expires, 12 months after approval. Deleted immediately when you delete your account |
| Temporary sign-in record created during school email verification | Deleted as soon as verification finishes (unless that email is itself a Lecta account) |
| Our admin action log (token grants, plan changes, approving or rejecting student verification and similar; includes the affected user's email and school name) | Deleted automatically after 1 year |
| Payment records (once paid features launch) | 5 years, as required by Korean e-commerce law, even if you delete your account |
When a retention period ends or the purpose is fulfilled, we delete the information without delay. Electronic files are deleted so they cannot be recovered.
You can delete your account in the app (Settings > Account & tokens > Delete account) or on the account deletion page.
4. Service providers and international transfers
Provider in Korea (once web payments launch): Toss Payments Co., Ltd. processes web payments from users in Korea. We send it the order number, product, amount and buyer email; Toss Payments collects card and other payment details directly. It keeps transaction records for the periods required by Korean law, such as the Electronic Financial Transactions Act.
We use or provide data to the providers below to run the service. All of them are companies outside Korea, so sharing data with them counts as an international transfer of personal information. Data is sent over encrypted connections (HTTPS) each time you use the related feature.
| Recipient (main country of processing) | Information sent | Purpose | When | Provider's retention |
|---|---|---|---|---|
| Google LLC (USA): Firebase, Google Cloud | Account information, device identifier hash, usage records, access logs, app crash reports, student verification requests (including the student ID photo), school email address (to send the verification email), payment records, push notification token, personal notices | Sign-in, sending the school email verification link, sending push notifications (Firebase Cloud Messaging), running the server, storing data, analysing app crashes (Firebase Crashlytics, USA) (database and server in the Seoul region; sign-in data may be processed and stored by Google outside Korea) | When you sign up and use the app | The periods in section 3 |
| Google LLC (mainly USA): Gemini API | Recording audio, transcript text, term hints, subject area, part of the previous sentence, lecture materials | Cloud speech recognition, AI correction and term extraction from lecture materials (default provider for all three, for users 18 or older) | When you use cloud transcription, AI correction or lecture materials | Paid tier: kept for up to 55 days to detect policy violations and meet legal obligations, then deleted. Only authorized staff may review it, and only for that purpose. Not used for AI training |
| DeepInfra, Inc. (USA) | Recording audio, term hints, part of the previous sentence | Cloud speech recognition (for users under 18 or without an age check record, or when the default provider has an outage) | When you use cloud transcription | Held in memory only while processing, not stored. Under the provider's policy it may be kept briefly for debugging. Not used for AI training |
| Google LLC (Singapore): Google Cloud (Cloud Run, Cloud Storage) | Recording audio, the resulting transcript | Cloud speech recognition (when users under 18 or without an age check record transcribe a whole recording file). Processed by a speech recognition model we run ourselves on Google Cloud; Google only provides storage and computing | When you use cloud transcription | Recording audio is deleted as soon as transcription finishes. The transcript is kept for up to 3 days so the app can fetch it, then deleted automatically. Not used for AI training |
| Groq, Inc. (mainly USA) | Recording audio, term hints, part of the previous sentence | Cloud speech recognition (backup, if we switch to it during another provider's outage) | When you use cloud transcription | Inputs (audio) and outputs (transcripts) are not retained (we have turned off data retention). The provider may keep usage metadata such as request counts. Not used for AI training. |
| Anthropic, PBC (mainly USA) | Transcript text, term hints, subject area (the field you chose), note template, lecture materials, terms extracted from lecture materials | AI correction and term extraction from lecture materials (for users under 18 or without an age check record, or when the default provider has an outage), AI notes (including lecture-material terms), translation | When you use those features | Normally deleted within 30 days. It may be kept longer for exceptions defined by the provider's policies, such as legal compliance or investigating policy violations. Not used for AI training. |
| Soniox, Inc. (USA) | Audio being recorded (real time), term hints, subject area | Instant captions (18 or older, default provider). Also cloud transcription if we switch cloud transcription to Soniox | When you turn on instant captions (and cloud transcription, if switched) | Real-time audio is handled only while processing and not stored. Audio and transcripts sent as files are kept on the provider's servers for up to 30 days, then deleted automatically. The provider's logs do not contain audio, transcripts or term hints. Not used for AI training |
| Google LLC (mainly USA): Gemini API | Audio being recorded (real time) | Instant captions (18 or older, when the default provider has an outage) | When you turn on instant captions | Paid tier: kept for up to 55 days to detect policy violations and meet legal obligations, then deleted. Only authorized staff may review it, and only for that purpose. Not used for AI training |
| Paddle.com Market Limited (UK) and its affiliates (USA and others) (once web payments launch) | Order number, product, amount, buyer email. Paddle collects payment details and billing country and postcode directly | Web payments from users outside Korea. Paddle is the merchant of record and handles payment, tax and refunds | When you pay on our website from outside Korea | As set out in Paddle's privacy policy and the transaction record periods required by law |
| Google LLC (USA): Google Drive | Recording files, transcripts, corrected transcripts, AI notes | Backup to your own Drive | When you turn on Drive backup | Until you delete it |
The AI providers (Google, Soniox, DeepInfra, Anthropic and Groq) process data mainly in the US and may also process it in other countries where they operate.
During the beta, the list of beta tester email addresses is also kept in repository settings at GitHub, Inc. (USA) and is deleted when the beta ends. If you sign in with Apple, Apple Inc. (USA) handles the sign-in. On Android, Google LLC processes the app integrity check (Play Integrity).
Through each provider's terms and data processing terms, we require them not to process the data for any purpose other than providing the service to us and to keep it secure. The AI providers do not use data received through their APIs for AI training (we use Google Gemini on the paid tier). When you delete your account, your sign-in data is also removed from Google's backups within 180 days.
How to refuse, and what happens: you can turn off cloud transcription in Settings > Recording & transcription > Fast cloud transcription, AI correction in Settings > AI correction & notes > Use AI correction, and Drive backup in Settings > Storage & backup > Google Drive backup. Instant captions are turned on and off with the Instant captions switch on the recording screen, and are off by default. AI notes send information when you create a note yourself, or each time a correction finishes if you turn on Settings > AI correction & notes > Create a note automatically after correction (off by default). If you don't want automatic notes, turn that option off. Translation sends information only when you run it yourself from the transcript screen. Once you turn a feature off or stop running it, the information it uses is no longer sent abroad from then on. Information already sent is handled under the provider retention described in the table above. If you do not agree to international transfers, you cannot sign in or use cloud features, but on-device transcription still works without signing in.
The app downloads its on-device speech recognition models from Hugging Face and GitHub. The app sends no other personal information for this; as with any download, your IP address is visible to those servers.
5. Your rights
You can ask at any time to see, correct or delete your personal information, to stop processing it, or to withdraw your consent. Email support@lectanote.com. We handle requests within 10 days and tell you the result. A legal representative or someone you authorize can also make a request for you.
- You can delete recordings and transcripts on your device yourself in the app (delete them from the recordings list).
- You can delete your account in the app (Settings > Account & tokens > Delete account) or on the account deletion page. Any remaining tokens are lost when you delete your account.
6. How we protect it
- All communication is encrypted with HTTPS.
- The access keys we use to run the service with outside providers are not built into the app; they are kept only in the server's secret storage. (If you enter your own key in the developer settings, it is stored only on your device and is not sent to our server.)
- We store only a one-way hash of the device identifier, not the original value.
- Only the server can write to the database, and the app can read only its own account's information.
- App Check blocks access from modified apps (applied to the Android app).
- Android's automatic app backup is turned off, so recordings are not copied off the device by accident.
7. Age
Lecta is not meant for children. In Korea you must be 14 or older to sign up. Outside Korea you must be 16 or older.
Age check: when you first open the app, you choose the year you were born, and the app decides which features you can use. If that year is on an age boundary, the app also asks whether your birthday is today or has already passed this year. We do not store the year. Only your age range (under the sign-up age / sign-up age to 17 / 18 or older) and the time of the check are stored on your device. Once you sign in, this range and time are also kept on our servers with your terms consent record. If you are under the sign-up age, you cannot sign in or use cloud features, but on-device recording and transcription still work. You can check your age again at any time in Settings > Account & tokens (if you chose the wrong year, or once you reach the sign-up age or turn 18). The new result replaces the old one, and if you are signed in, the record on our servers is updated too. If an existing account can no longer sign in because the check shows you are under the sign-up age, you can ask us to delete it on the account deletion page.
The Gemini API and Soniox are only for people 18 or older: the Gemini API and Soniox terms allow only people 18 or older to use the service. So only users whose age check shows 18 or older can turn on instant captions, and our server refuses to open an instant caption connection without an age check record showing 18 or older. Cloud transcription checks the same record: only audio from users 18 or older goes to the Gemini API; audio from everyone else goes to a speech recognition model we run ourselves on Google Cloud (Singapore) or to DeepInfra. AI correction works the same way: only transcripts from users 18 or older go to the Gemini API; transcripts from everyone else go to Anthropic (Claude). Recording and on-device transcription work at any age. Cloud transcription, AI correction and other cloud features are available from the sign-up age.
8. Contact
- Privacy contact: Lecta (the operator)
- Email: support@lectanote.com
Users in Korea can also contact the Personal Information Infringement Report Center (privacy.kisa.or.kr, 118) or the Personal Information Dispute Mediation Committee (kopico.go.kr, 1833-6972).
9. If you are in the EEA, UK or Switzerland
Controller: Lecta, the operator of the app, is the controller of your personal information. Contact: support@lectanote.com.
Legal bases:
- Contract: to provide the service you signed up for: your account, sign-in, tokens and plans.
- Consent: for cloud processing of your audio and text (cloud transcription, AI correction, AI notes, translation, instant captions, Drive backup) and the international transfers this involves. You give consent on the consent screen and when you turn these features on. You can withdraw it at any time by turning them off (see "How to refuse" in section 4), or by deleting your account (sign-in data is sent abroad as part of signing in). Withdrawing stops future transfers; data already sent is kept by the provider as described in section 4.
- Legitimate interests: to prevent abuse (such as repeat free-token claims), keep the service secure, check the app is genuine and fix errors.
Your rights: you have the right to access your personal information, to have it corrected (rectification) or deleted (erasure), to restrict its processing, to receive it in a portable format (portability), to object to processing based on legitimate interests, and to withdraw consent at any time. Withdrawing consent does not affect processing that happened before. To use these rights, email us.
Complaints: you have the right to complain to a data protection supervisory authority, in particular in the country where you live or work, or where you think a breach happened.
International transfers: our database and server are in South Korea (Google Cloud, Seoul region); sign-in data may be processed and stored by Google outside Korea. The European Commission and the UK have both recognized South Korea as providing an adequate level of data protection (adequacy decisions). Transfers to the providers listed in section 4, which process data mainly in the US and may also process it in other countries where they operate, rely on their data processing terms, including standard contractual clauses where those apply.
10. If you are in the United States
No sale or sharing: we do not sell your personal information, and we do not share it for cross-context behavioral advertising. The app includes Google's advertising SDK (AdMob) for ads that may be added later, but the SDK is not started unless ads are turned on, and ads are currently off. The app contains no analytics SDK. It uses Firebase Crashlytics only to report app crashes to us (section 1).
California residents: you have the right to know what personal information we collect, use and disclose (described in sections 1 to 4), to request deletion, to request correction, and to opt out of the sale or sharing of personal information (we do neither). We use sensitive personal information only to provide the service you ask for. We will not treat you differently for using these rights. To make a request, email us; you may also use an authorized agent. We may need to confirm your identity first.
Do Not Track: the app does not track you across other apps or websites, so it does not respond differently to Do Not Track signals.
Children: Lecta is not for anyone under 16 outside Korea (see section 7); in particular, we do not knowingly collect personal information from children under 13. If you believe a child has given us information, contact us and we will delete it.
11. If you are in Japan
Provision to third parties in foreign countries (Act on the Protection of Personal Information, Article 28): when you use the features described in section 4, your personal information is provided to recipients in the following countries:
- South Korea: our database and server (Google Cloud, Seoul region).
- Singapore: Google Cloud (Cloud Run, Cloud Storage), where we run our own speech recognition model.
- United Kingdom (once web payments launch): Paddle.com Market Limited.
- United States: Google LLC, Soniox, Inc., DeepInfra, Inc., Groq, Inc., Anthropic, PBC, Apple Inc. and GitHub, Inc.
- Other countries where the provider operates: the AI providers (Google, Soniox, DeepInfra, Anthropic and Groq) process data mainly in the US and may also process it in other countries where they operate. Google may also process and store sign-in data outside Korea.
Personal information protection systems in those countries (summary):
- South Korea has a comprehensive law, the Personal Information Protection Act, overseen by an independent regulator, the Personal Information Protection Commission (PIPC).
- The United States has no comprehensive federal privacy law. Sector-specific federal laws and state laws apply, such as the Federal Trade Commission Act and the California Consumer Privacy Act (CCPA).
- Singapore has a comprehensive law, the Personal Data Protection Act 2012 (PDPA), overseen by the Personal Data Protection Commission (PDPC).
- The United Kingdom has a comprehensive law, the UK GDPR and the Data Protection Act 2018, overseen by an independent regulator, the Information Commissioner's Office (ICO).
- For other countries where providers operate, the system depends on the country. You can find information on foreign systems on the website of Japan's Personal Information Protection Commission (PPC).
Measures taken by recipients: each provider processes the data under contractual data processing terms; the AI providers do not use it for AI training (we use Google Gemini on the paid tier); and data is kept as described in section 4.
Our seller information under the Act on Specified Commercial Transactions is on 特定商取引法に基づく表記.
12. Changes to this policy
If we change this policy, we will announce it in the app 7 days before the change takes effect, or 30 days before for important changes to your rights.