Privacy Policy
Nibab (formerly Amharic TTS) is an app that reads Amharic text aloud. This policy explains what information the app handles, where it goes, and who can see it.
In short
- We, the developer, don't collect, see or store your text, audio, photos or handwriting. We have no servers for the app.
- Text to speech and image to text run entirely on your device and work without an internet connection.
- The Handwriting feature uses Google ML Kit. Downloading its model needs the internet, and ML Kit sends Google limited technical diagnostics. Your handwriting is not sent.
- No account, no ads, and no analytics or tracking by us.
Information we collect
We don't collect any personal information. The app has no account or sign-in, and we don't run analytics, advertising or tracking. Nothing you type, record, write or photograph in the app is sent to us.
What happens on your device
- Text to speech: the text you enter is turned into speech on your device. The text and audio are kept in the app's memory while you use it. When you tap Share, the app creates a temporary WAV file and hands it to the iOS share sheet; you decide where it goes.
- Image to text: you choose a photo with the iOS photo picker, so the app only receives the photo you pick, not your photo library. Text in the photo is recognized on your device. The app doesn't upload or keep the photo.
- Handwriting: after the handwriting model is downloaded, your strokes are recognized on your device and aren't sent anywhere.
Google ML Kit (Handwriting feature)
The Handwriting feature uses Google ML Kit Digital Ink Recognition.
- Model download: the first time you tap Download Amharic model, the app downloads the Amharic handwriting model (about 14 MB) from Google's servers over Wi-Fi or mobile data. As with any internet request, Google's servers receive your device's IP address.
- Diagnostics: according to Google's ML Kit data disclosure, ML Kit sends Google device information (such as manufacturer, model and OS version), the app's bundle ID and version, a per-installation identifier that is not intended to identify you or your device, performance metrics, API configuration, event types and error codes (such as model downloads), and the configured recognition language. Google uses this for diagnostics and usage analytics.
- ML Kit does not send your handwriting, the recognized text, or any text you type.
Google's handling of this information is covered by the Google Privacy Policy.
Other components
Speech is generated with Meta's open MMS (Massively Multilingual Speech) model running in ONNX Runtime, and image text is recognized with Tesseract OCR. These run on your device and don't send data anywhere. The app has no other third-party services, analytics or advertising networks.
Data retention and deletion
Everything the app keeps is on your device, and deleting the app removes it, including the downloaded handwriting model. Audio files you shared or saved elsewhere stay wherever you put them. We hold no data about you, so there is nothing for us to delete.
Children's privacy
Nibab is suitable for all ages. We don't knowingly collect personal information from anyone, including children under 13. The Google ML Kit diagnostics described above apply to all users of the Handwriting feature.
Your choices
- You can use text to speech and image to text without the Handwriting feature and without an internet connection.
- You choose which photos the app receives, and where shared audio goes.
Changes to this policy
If we change this policy, we'll post the new version at nibab.app/privacy with a new effective date.
Contact
Questions about this policy? Email [email protected].