SpeakCoveYour files stay yours.
The web reader opens your file in your browser and reads it aloud right there: no sign-up, no install, no upload. The app for iPhone and iPad needs no account either, and carries no analytics, tracking or ad SDKs.
Voices come in. Your file stays.
The reader’s tab on speakcove.com, with what it downloads to read aloud: the voice model and the engines that run it. Your file is opened and read inside the tab and never goes the other way.
Open the reader.
It is a page on speakcove.com. There is no account to make and nothing to install.
Drop in your file.
Your browser hands it to the page, and the text is pulled out right there in the tab. Nothing is uploaded.
Press play.
The voices download once, about 100 MB, then read your text aloud on your own device.
What goes where
What the page fetches, and what it never sends.
Opening a file in the web reader works much like opening it in an app on your computer. Your browser reads the file from your disk and hands it to the page, and the page pulls the text out right there in the tab. Pages that are only scans go through text recognition in the tab as well. No copy of the file goes to a server, ours or anyone else’s.1
The voices are the part that needs a download. The first time you press play, your browser fetches SpeakCove’s voice model from speakcove.com, about 100 MB, and keeps it, so later visits start in a second or two. The engines that run the voices, open PDFs and read scans come from the jsDelivr CDN, which sees your IP address like any web server. Those downloads are the voices and the code that runs them. Your file is not among them, and it never travels the other way.
That makes the web reader a sensible place for the things you would rather not upload anywhere: a contract, a letter from your doctor, a draft that isn’t finished. The reader keeps no copy. Close the tab and the text is gone; only the voices stay in your browser for next time.
In the app the same holds, without a browser around it. There is no account, so SpeakCove never knows who you are. Your files stay on your phone or tablet, and scanning, text recognition and the voices all run on the device, in airplane mode too. The app carries no analytics, crash reporting, tracking or ad SDKs.2
It goes online only when you ask it to: to search the free books and papers, to download the one you pick from its source, such as Project Gutenberg or arXiv, to sync through iCloud, and for App Store purchases. Search terms pass through our own caching server, which keeps no log of requests.
One limit is worth knowing. SpeakCove’s voices need WebGPU, which current versions of Chrome, Edge and Safari have. In a browser without it, the reader falls back to the built-in voice of your browser or device, and where that voice runs is up to them: some built-in voices work online. The reader tells you when that happens.
- Account
- Not needed, on the web or in the app. SpeakCove doesn’t know who you are.
- Install
- None for the web reader. It is a web page.
- Your file
- Opened and read in the browser tab. Never uploaded.
- Scanned pages
- Recognized in the tab. The recognition models download once from speakcove.com, up to 31 MB.
- Web voices
- All 10 run on your device with WebGPU. Downloaded once from speakcove.com, about 100 MB, and kept by your browser.
- Engines
- ONNX Runtime and PDF.js, from the jsDelivr CDN.
- Browsers
- Current Chrome, Edge and Safari. Elsewhere it reads with the built-in voice of your browser or device.
- Website analytics
- Google Analytics, cookies only if you accept them. It notes play, voice, language and file type. Never file names or text.
- Files in the app
- On your phone or tablet. With iCloud on, also in your own iCloud account.
- Camera scans
- Text recognized on the device.
- App voices
- On the device. They work in airplane mode.
- App analytics
- None. No crash reporting, tracking or ad SDKs.
- App goes online for
- Searching free books and papers, downloading them from their sources, iCloud sync, and App Store purchases.
- Devices
- iPhone and iPad with iOS 26 or later.
- In full
- Each online feature, and exactly what it sends, is listed in the Privacy Policy.
Is my file uploaded when I use the web reader?
No. Your browser opens the file and the page reads it in the tab. The page downloads the voices and the engines that run them; nothing of your file goes the other way.
Do I need an account?
No. The web reader has no sign-up, and the app has no account at all, so SpeakCove never knows who you are.
Does speakcove.com use cookies?
The website uses Google Analytics to see which pages are useful. Its cookies are set only if you accept them in the cookie banner, and you can change your choice at any time with “Cookie settings” at the bottom of every page. Your file and your text are never part of it.
Does the SpeakCove app track me?
No. On iPhone and iPad the app carries no analytics, crash reporting, tracking or ad SDKs, and your documents and scans stay on your device.
Try it with your own file.
Open the reader, drop in a PDF and press play. It stays in your browser. The app for iPhone and iPad keeps your files on the device too.
Scans and paper
Scanned PDFs and photos of paper are recognized, then read aloud.