Privacy Policy for Audio Transcriber

1. Information the App accesses and processes

The App accesses only content that the user explicitly selects or shares with it, together with the settings needed to provide its features. Depending on how the App is used, this may include:

• Audio files and the audio tracks of compatible video files selected or shared by the user.
• Transcription text generated from that content and an optional initial prompt entered by the user.
• User-assigned transcription names and operational metadata, such as request time, audio duration, processing time, selected language, model, and transcription method.
• App preferences, including interface and transcription language, model selection, processing options, and whether audio should be retained in history.
• Downloaded Whisper and Silero model files required for local speech processing.

The App does not create user accounts and does not ask for a person’s name, postal address, phone number, payment information, contacts, precise location, or advertising identifier.

2. How information is used

The selected audio or video audio track is copied to the App’s private storage so it can be queued, processed, and transcribed on the device. The App uses the generated text and related metadata to display progress and maintain a searchable local history. If “Save audio in history” is enabled, a private copy of the source audio is retained so the user can replay it from the corresponding history entry.

The App may display local Android notifications for model downloads, transcription progress, completion, copying a result, and cancellation. Notification permission is optional; refusing it does not prevent local transcription.

The developer does not use App data for analytics, profiling, advertising, marketing, or tracking.

3. Local processing and network use

Speech recognition and voice-activity detection run entirely on the device using whisper.cpp and Silero VAD. The App does not transmit audio, video, prompts, or transcription results over the internet.

The App uses internet access only to download speech-recognition and voice-activity-detection model files from Hugging Face. During a model download, Hugging Face and its infrastructure providers may receive standard connection information, such as the device’s IP address, request time, requested file, and HTTP metadata. This information is handled by Hugging Face under its Privacy Policy. The developer does not receive model-download logs from Hugging Face.

4. Data sharing and sale

The developer does not collect, sell, rent, or share the user’s audio, video, prompts, transcriptions, history, or App settings with third parties.

If the user chooses Android’s Share action for transcription text, the selected text is sent to the destination app chosen by the user. That transfer is initiated and controlled by the user and is then governed by the destination app’s privacy practices.

Limited technical connection information is sent directly to Hugging Face when the user or the App downloads a required model, as described above. Android backup and device-transfer services may also process selected local data as described in the next section.

5. Storage, backup, and device transfer

App data is stored within Android’s private app storage, which is isolated from other apps under Android’s security model unless the user deliberately shares content.

• Remote Android backup: only App settings, including the optional initial prompt, are eligible for Android’s cloud backup. Audio, transcription history, pending queue data, downloaded models, temporary files, and cache are excluded.
• Direct device transfer on Android 12 and later: App settings and textual transcription history with its metadata may be transferred directly to a new device by Android. Audio files are excluded.
• Android 11 and earlier: only App settings are eligible for backup or transfer; transcription history and audio are excluded.

Android backup and device-transfer availability depends on the user’s device and system settings. When Google provides the backup service, the transferred information is handled under the Google Privacy Policy. The developer cannot view or access the contents of a user’s Android backup.

6. Retention and deletion

Pending audio is retained locally while it is queued or being processed. Temporary processing files are removed when they are no longer required. Completed transcription text and its metadata remain in local history until the user deletes them. Saved history audio remains until its history entry or the audio itself is deleted.

Users can control and delete local data directly in the App:

• Delete an individual history entry, which also deletes its saved audio.
• Use Settings → Storage management → Delete all transcriptions and audios to remove all transcription history and all active, pending, and saved audio while keeping downloaded models.
• Use Delete all audios to remove active, pending, and saved audio while keeping transcription text and models.
• Use Delete all models to remove downloaded Whisper and Silero models while keeping transcription text and saved audio.
• Clear the App’s storage or uninstall the App to remove its local data from the device, subject to any settings already retained by the user’s Android backup service.

Because the App has no user accounts and the developer does not receive or store user content on developer-controlled servers, there is no server-side account or user-content database for the developer to delete. Questions about deletion may be sent to the privacy contact listed below.

7. Security

The App minimizes exposure by processing audio and transcription content locally, storing data in Android’s private app storage, excluding sensitive content from cloud backup, and using encrypted HTTPS connections for model downloads. Downloaded models are checked against their expected size and SHA-256 digest before use.

No storage or transmission method can be guaranteed to be completely secure. Users should keep their device and Android version updated, use an appropriate screen lock, and avoid sharing sensitive transcription text with untrusted apps or recipients.

8. Children’s privacy

The App is not directed to children under 13 and the developer does not knowingly collect personal information from children. Because transcription content remains on the device and is not collected by the developer, the developer cannot inspect the age or content of App users.

9. Changes to this Privacy Policy

This Privacy Policy may be updated when the App’s features, data practices, or applicable requirements change. The “Last updated” date at the top of this page will identify the latest revision. Material changes will be reflected in the policy made available with the App and on its Google Play listing.

10. Contact

For privacy questions or requests concerning Audio Transcriber, contact:

overcoders
Email: [email protected]
Website: https://neftali.com.ar/
Country: Argentina