Convert an EPUB without uploading it
The book is read off your disk and narrated in the tab. Nothing is sent anywhere, it works with the wifi switched off, and you can watch the network stay silent while it runs.
Local narration unit
Idle
Drop an EPUB here
M4B
With chapters
0
Bytes uploaded
11
Voices
Read aloud in this tab. Never uploaded. Here's why that's true.
book.epub
- —
- Chapters
- —
- Audio
- —
- On this machine
- —
- Output
Reading the book…
—
Close the tab if you want. We keep your place, and picking the same file again carries on from where it stopped.
Paused
Your place is kept.
Your audiobook is ready
Getting it onto a phone
- iPhone. AirDrop it, or copy it into the Books app. Chapters and your position work straight away.
- Android. Copy it across and open it in Smart AudioBook Player, Voice or Listen.
- Plex or Audiobookshelf. Drop it in the library folder.
That didn't work
- 0 bytes
- Uploaded, ever
- Offline
- After the first run
- No account
- Nothing to sign up to
Check it rather than believe it
Every site on this search claims to respect your privacy, and the claim costs nothing to make. This one is unusual only in being checkable in about thirty seconds:
- Open devtools — F12, or Cmd+Option+I on a Mac.
- Go to the Network tab and leave it open.
- Drop a book on the page and start a conversion.
You will see the page load, then the voice model download once — that is the only sizeable request, and it comes from our own storage bucket. Then nothing. The list stops growing while the conversion runs, for however many hours it runs, because there is nowhere for a request to go.
The stronger version of the same test: load the page, turn off your wifi, and convert. It finishes.
Why this is worth caring about
A book says a lot about the person reading it. Uploading one to a converter means handing a stranger a copy of something you chose, at a moment they can timestamp, from an address they can log. Most of the time nothing comes of it. The files sit on a disk in a datacentre until a retention policy nobody has read deletes them, or doesn't.
It is also, quite often, someone else's copyright sitting on a third party's server, which is a different problem wearing the same coat.
What we actually cannot do
Not a privacy policy, a description of the architecture. There is no application server: the site is static files served from a CDN. There is no upload endpoint to receive a book, no database to put one in and no queue to hold one. Nothing was switched off to make this true — it was never built, because the conversion happens on your machine and there was nothing for a server to do.
The one thing we do fetch is the voice model, from our own bucket, and it is the same file for everybody. It tells us nothing except that somebody, somewhere, is about to listen to something.
Questions
How can I be sure nothing is uploaded?
Check it. Open devtools before you convert, go to the Network tab, and watch. You will see the page load and the voice model download once. After that the network goes quiet and stays quiet for the whole conversion, however long it runs. It is the kind of claim that is easy to make and easy to verify, so verify it.
Does it work offline?
After the first run, yes. The voice is about 63 MB and the browser keeps it. Load the page once with a connection, then turn the wifi off entirely and convert. On a plane, on a train, on a laptop that has never been on your home network — it does not care.
What about the file I download at the end?
It is assembled in the browser and saved straight to your downloads folder. It does not pass through us on the way, because there is no us to pass through — the site is static files on a CDN with no application server behind them.
Do you keep logs, analytics or cookies?
No analytics, no cookies, no tracking pixels and no third-party scripts of any kind. The content security policy on every page refuses third-party scripts outright, and a browser test in our build pipeline fails if any manage to run. That check exists because our own host once tried to inject one.
Is my book stored in the browser?
The audio is, while it is being made — that is what lets you close the tab and resume. It lives in the browser's private storage for this site, which no other site can read, and it is deleted the moment the finished file is assembled. Clearing site data removes it immediately.
Why do other converters need to upload?
Because they run the speech model on their own hardware, which is faster and costs them money. Uploading is not a design flaw on their part, it is how their product works. The trade here is the opposite one: slower, and the book stays where it is.