VoxReader
Privacy Policy
VoxReader collects nothing about you. There is no account, no analytics, no advertising and no tracking. This page exists to say exactly what does leave your device, when, and at whose request.
Effective 20 September 2026 · Last updated 20 September 2026
The short version
Your library, your reading positions, your listening history and your settings stay on your device. The developer of VoxReader operates no server, receives no copy of anything you read, and has no way to identify you.
The app does make network requests — but only to addresses you chose, and only for the feature you asked for. Every one of them is listed below.
What VoxReader collects
Nothing. In App Store terms, this app's data collection answer is Data Not Collected. VoxReader has no account system, no sign-in, no analytics SDK, no crash reporter that phones home, no advertising identifier and no third-party tracking of any kind. Nothing you import, read or listen to is transmitted to the developer, because there is no service on the other end to transmit it to.
What stays on your device
- Everything you import — text, HTML, EPUB, PDF, RTF, FB2 and DOCX files, and web pages you share into the app — together with the normalised text the app builds from it.
- Your library, folders, bookmarks and reading positions.
- Your listening history and playback statistics.
- Your settings, including which voices you have chosen.
- Any synthesised audio the app caches.
These live in the app's own storage. If you have iCloud backup enabled for your device, they may be included in your own encrypted device backup — that is Apple's mechanism and your account, not the developer's.
When VoxReader uses the network
VoxReader is an offline reader by design: reading a book aloud with the device's own voices requires no network at all. The app connects only in these situations, each of which you initiate.
| What you do | Where the request goes | What is sent |
|---|---|---|
| Add or refresh an RSS / Atom / JSON feed | The feed's own address, which you entered | An ordinary HTTP request for that feed |
| Browse or download a public-domain book | The catalogue you picked (for example Project Gutenberg or Standard Ebooks) | A search or download request |
| Import a web page | That page's address | A request for the page, so its text can be extracted on your device |
| Open a link in the in-app browser | The site you tapped | Whatever a normal browser sends; VoxReader adds nothing |
Those servers are operated by other people, and their own privacy policies govern what
they log — typically an IP address and a request line, as any web server does. VoxReader
sends nothing that identifies you with these requests. Requests to book catalogues
identify the app, as a polite client should:
VoxReader/1.0.0 (+mailto:voxreaderapp@gmail.com) — the app's name, its version,
and an address to complain to. There is no user identifier, no device identifier and no
account in it.
The two online voices
Most voices in VoxReader run entirely on your device. Two are different: they synthesise speech on someone else's servers, which means the sentence being read has to be sent there. Both are clearly marked Online in the voice picker, and neither is ever selected for you automatically — a book will never start using one because the app matched a language.
Google Translate voices — free, no account
These are the voices behind the speaker button on the Google Translate web page.
- They are listed by default in the voice picker. You can turn them off in Settings, and doing so stops a book that is already playing with one.
- They are never chosen automatically. If you never open the voice picker and pick one, nothing is ever sent to Google.
- When you do use one, each sentence is sent as an HTTPS request to
translate.google.<tld>/translate_tts, and the spoken audio comes back. - There is no account, no API key and no identifier. Google receives the sentence and the network request itself; what Google logs is governed by Google's privacy policy.
- The decoded audio is cached on your device, so re-reading a passage does not send it again.
OpenAI voices — off by default, your own account
- They are off by default and require two things before they can be used: switching them on in Settings, and pasting an OpenAI API key you created on your own OpenAI account.
- With one selected, each sentence is sent to
api.openai.comunder your key, and the usage is billed to your OpenAI account. Nothing is proxied through the developer, who sells nothing here and takes no part in that billing. - Your API key is stored in the device Keychain, is excluded from backups, and is sent to no one but OpenAI. The app's own screens only ever show it masked.
- The synthesised audio is cached on your device, so a book you have paid to synthesise once plays again — including offline — without being paid for twice.
- What OpenAI does with text sent under your key is governed by your agreement with OpenAI.
Using either voice means you have chosen to send the text you are reading to that company. That is your device talking to a third party at your instruction; it is not collection by the developer of VoxReader, who never sees it.
Children
VoxReader is a general-audience reading app. It does not knowingly collect any information from anyone, of any age, because it does not collect information at all.
Changes to this policy
If a future version of VoxReader adds anything that sends data anywhere, this page will say so before that version ships, and the date at the top will change.
Contact
Questions about this policy, or about anything the app does with your data: voxreaderapp@gmail.com.