Privacy Policy

Last Updated: 14/09/2026

1. Overview

Matt is designed with privacy in mind. You choose the transcription engine: with on-device transcription (Pro) the audio is processed entirely on your iPhone by Apple's speech frameworks and never leaves it; with AI transcription (Premium) the audio is processed by a cloud AI provider (AssemblyAI) over an encrypted connection and is not stored. We do not require an account: no email, name, or personal information is collected.

Two features send data to third-party AI services, and each one is used only after you have given your explicit consent inside the app:

• AI Transcription (Premium, formerly called "Advanced Transcription") sends the audio you select to AssemblyAI to transcribe it, detect the language and recognize the different speakers.
• Summaries and automatic titles (available in both Pro and Premium) send the transcribed text to our server (Supabase), which forwards it to OpenAI to generate the summary or the title.

The first time you use one of these features, the app shows you an alert explaining what will be sent and to whom, and asks you to accept. If you decline, the feature is not performed and no audio or transcribed text is sent to any third-party AI service. The sections below describe exactly which data leaves your device, when, why, and to whom.

2. Key Privacy Points

• On-device transcription (Pro): audio, and its translation, are processed entirely on your device and are never sent anywhere. Works offline.
• AI Transcription (Premium): if you use it, the selected audio is sent to AssemblyAI to transcribe it, detect the language and recognize speakers, only after you consent. It is not stored and not linked to any personal identity.
• Summaries and automatic titles: after you consent, the transcribed text is sent to our server (Supabase) and forwarded to OpenAI — only when you explicitly request a summary, or, for automatic title generation, only when you save the audio inside the app. It is not linked to any personal identity.
• Consent is requested in the app the first time you use each of these features, with an alert stating what is sent and to whom. If you decline, nothing is sent.
• No user accounts — no email, name, or personal information collected.
• Firebase Analytics collects only anonymous usage data.
• RevenueCat handles payments — purchase data is processed through the Apple App Store.
• All your data stays on your device — no proprietary cloud backups.

3. RevenueCat (Subscription Management)

Purpose: manages in-app subscriptions and purchases.

Data transmitted: subscription status (free/Pro/Premium), purchased plan type (weekly/yearly), free trial eligibility, purchase history, billing issues.

User identifier: anonymous (generated by RevenueCat).

See the RevenueCat Privacy Policy for details.

4. Firebase Analytics (Google)

Purpose: tracks app usage events to improve the product.

Data transmitted: anonymous behavioral events with no personally identifiable information.

Tracked events include:

• Onboarding completed / step reached
• Paywall shown (with reason), dismissed, plan selected
• Purchase started / completed
• Transcription completed (duration in seconds + language)
• Share Extension opened / closed
• Actions used (copy, save, translate, summarize)

Identifier: anonymous Firebase Instance ID.

See the Google Privacy Policy for details.

5. Supabase + OpenAI (Summaries and Auto Titles)

Purpose: generates summaries and automatic titles for your transcriptions, in both the Pro and Premium plans.

Consent: the first time you request a summary, and the first time a title is generated, the app shows an alert explaining that the transcribed text will be sent to our server and to OpenAI, and asks you to accept. Until you accept, no transcribed text is sent to Supabase or OpenAI. If you decline, the feature is not performed.

When it is used (after consent):

• Summaries: only when you explicitly request a summary.
• Automatic title generation: only when you save the audio inside the app. If you do not save the audio, no title generation is performed and no text is sent.

Data sent to Supabase (intermediate server):

• Transcribed text (max ~120,000 characters for summaries, max 1,500 for titles)
• Transcription language
• Requested level of detail

Supabase then forwards the text to OpenAI for processing.

Authentication: anonymous (no account, no email).

Note: the transcribed text is NOT associated with any personal identifier.

See the OpenAI Privacy Policy for details.

6. AssemblyAI (AI Transcription)

Purpose: cloud speech-to-text with automatic language detection (including mixed languages) and speaker recognition.

When it is used: only when you use AI Transcription (Premium). The first time you use it, the app shows an alert explaining that the selected audio will be sent to AssemblyAI, and asks you to accept. If you use on-device transcription instead, or you decline the alert, no audio is sent to AssemblyAI.

Data sent: the audio you selected (for video files, only the audio track extracted on your device) and, where applicable, the transcription language.

Retention: the audio is processed over an encrypted connection for transcription only, is not stored after processing and is not used to train models.

Authentication: anonymous (no account, no email). The audio is not associated with any personal identifier.

See the AssemblyAI Privacy Policy for details.

7. Apple Speech Framework (On-device Transcription and Translation, Pro plan)

Processing: entirely on-device.

When you use on-device transcription (Pro), no audio or text data is transmitted to external servers during transcription or translation, and both work offline.

AI Transcription (Premium) instead uses AssemblyAI, described above.

8. Apple Intelligence / FoundationModels (Alternative Summaries)

Processing: entirely on-device on iOS 26+.

Used as a local fallback when there is no internet connection or OpenAI is unavailable.

No data is transmitted.

9. How We Collect Data

Matt does not collect data in the background and does not build a profile of you. Data is collected only as a direct result of an action you take in the app:

• Audio is captured by the microphone when you start a recording, or imported when you share a voice message to the app from WhatsApp, Telegram or another app, or when you import an audio or video file. It leaves your device only for AI Transcription (Premium), and only after your consent.
• Transcribed text is produced on your device from that audio. It leaves your device only when you request a summary or save an audio (title generation), and only after your consent.
• Usage events are collected automatically and anonymously by Firebase Analytics while you use the app (see section 4).
• Purchase data is received from the Apple App Store through RevenueCat when you buy or restore a subscription (see section 3).

We never collect your name, email address, phone number, contacts, or location, and we do not sell or rent any data to anyone.

10. Third-Party Protection

We share data only with the providers listed in this policy, and only for the purposes described here. Each of them is bound by contractual terms and by its own published privacy policy to provide a level of protection for your data that is the same as or equivalent to the one described in this document. In particular, all of them are required to:

• process the data only to deliver the service requested, and for no other purpose;
• not use your audio or transcribed text to train their AI models;
• not sell, rent, or disclose the data to further third parties;
• apply encryption in transit and appropriate security measures;
• retain the data no longer than necessary to deliver the service.

The providers we use are: AssemblyAI (AI transcription), OpenAI (summaries and titles, via its API, which is excluded from model training by default), Supabase (intermediate server), Google Firebase (anonymous analytics) and RevenueCat (subscription management).

11. Local Data (Never Transmitted)

• All transcriptions, audio recordings, summaries, and translations are saved only on your device (SwiftData).
• Preferences, language settings, and Pro/Premium status are stored locally (UserDefaults).

12. Permissions Requested

• Microphone — to record audio.
• Speech Recognition — to transcribe audio.
• Notifications — optional, used for feedback reminders.

13. User Rights

Since Matt does not collect personally identifiable data and does not require an account, there are no personal records to access, correct, or delete on our side. You can delete any transcription, recording, summary, or translation directly in the app, which removes it permanently from your device. Since the AI features are optional, you can simply stop using them and no further data will be sent. Anonymous data sent to RevenueCat, Firebase, Supabase, OpenAI, or AssemblyAI is governed by their respective privacy policies linked above.

14. Disclaimer

This is an UNOFFICIAL tool that can be used with WhatsApp and Telegram. We are NOT affiliated, associated, authorized, endorsed by, or in any way officially connected with WhatsApp LLC or Telegram, or any of its subsidiaries or affiliates. Usage falls within guidelines of "fair use".

15. Contact Us

If you have any questions or concerns about this Privacy Policy, please contact us at albosapps@gmail.com.