
Privacy policy
The short version
Inside Voice does not require an account. It does not sell data, run ads, or embed third-party analytics. Your books and reading progress stay on your device. Optional features (cloud AI keys, microphone transcription, neural voice download) only run after you turn them on, and only send what that feature needs.
This policy describes what Inside Voice does with data on your device, and what leaves the device when you choose to switch on an optional feature.
Data stored on your device
- Books you import (EPUB or PDF), copied into the app's storage so moving the original file does not break the library.
- Reading position, themes, voice choices, speed-reading settings, and similar preferences.
- Rendered chapter audio and word timings, cached so a chapter does not have to be rebuilt every time you open it.
- Optional API keys you enter for Gemini, Azure OpenAI, or Azure Speech, kept in the Keychain on your device.
- Your own voice, if you record one. The ten second recording is written to the app's temporary folder and the voice built from it is kept in the app's own storage. Neither is uploaded, and neither is put anywhere other apps or the Files app can see. Deleting the voice deletes both.
- Notes, quotes and the Board you build from a book, and how much of each book you have read.
- Prepared audiobooks you asked the app to keep, and audiobooks you export, which go to the app's Documents folder so you can reach them in Files.
The App Store privacy label for this build declares no tracking, and no collected data types sold or used for tracking. The app reads preferences, file timestamps, and system boot time only for ordinary on-device behaviour: settings, cache freshness, and timing.
Microphone
The microphone is used for two things, each only when you start it.
- Recording your own voice so books can be read back in it. That recording never leaves the device: the voice is built on the phone, and both the recording and the voice are deleted when you delete the voice.
- Speaking a recall answer instead of typing it. This device's own speech recognition handles it, and nothing leaves the phone. If you have set an Azure Speech key and region in Settings, a recording it cannot handle is sent to Microsoft Azure Speech for short transcription and discarded by the app once the text comes back.
Typing always works without the microphone, and the app reads books with no microphone permission at all.
Optional network features
None of these are required to read a book aloud with system voices.
- On device (default). Nothing leaves the phone. Available when Apple Intelligence is ready.
- Gemini. Sends text you have already read, plus your recall or chat messages, to Google Gemini using an API key you paste. Only if you select Gemini.
- Azure OpenAI. The same kinds of text, sent to your own Azure OpenAI resource. Only if you configure endpoint, deployment, and key.
- Azure Speech. Sends short microphone recordings you start to Microsoft Speech, using your key and region. Only for voice input you trigger.
- Neural voice install. Downloads model weights (via the neural voice package, from Hugging Face) only if you tap to install the optional voice.
Keys never leave the Keychain except as authorization headers on requests you start. The app does not operate a backend that stores your books or your chat history.
Tips
The app is free and nothing in it is paid for. There is an optional tip jar at the foot of the reading statistics screen, and it unlocks nothing: the app is identical before and after.
The tip jar is handled by RevenueCat, on top of Apple's own In-App Purchase. It is not started until you open that screen, so a reader who never goes there never contacts it. When it does run, RevenueCat receives an anonymous identifier it generates for the install and the purchase result. It never receives your name, your books, or anything you have read. Apple handles the payment itself and this app never sees your payment details.
Children
Inside Voice is not directed at children under 13. Do not use the optional cloud features to send a child's personal information.
Third-party services
If you enable Gemini, Azure OpenAI, Azure Speech, the neural voice download, or leave a tip, those providers' own terms and privacy policies also apply to the data that reaches them. Inside Voice does not control how those providers retain logs on their side.
The app sends a passage only to the provider you chose in Settings. If that provider fails, the app reports the failure; it does not quietly send the same passage to a different one.
Changes
If this policy changes in a way that affects how the app handles data, the effective date at the top will update, and the App Store listing privacy URL will keep pointing here: https://bassamkamal.dev/insidevoice/privacy.html
Contact
Questions about privacy: bassamkamal.py@gmail.com. For help with the app, see Support.