PDF to Audiobook

A long report or an out-of-copyright book is easier to finish on a walk than on a screen. PDF to Audiobook is built for that. You drop in an English PDF with selectable text, up to 20 MiB, or paste your own text. The site extracts the words in your browser and shows them to you before anything is converted, which matters because PDFs are laid out for the page, not for listening. Running headers, page numbers and footnotes get read aloud if you leave them in, so the workspace lets you decide what stays, then rename, reorder, split or merge sections into the chapters you want to hear. Each conversion can cover up to 200 sections and 600,000 characters, and finished chapters download as MP3 files.

There are three preset voices, and you can play samples of each reading the same passage without an account. Two use ElevenLabs Eleven v3, a lower British storytelling voice and a softer conversational one, and the third, meant for reports and study material, uses MiniMax Speech 2.8 HD. The audio is AI-generated, and the site doesn't offer voice cloning.

Scanned PDFs need OCR in another tool first, and password-protected or copy-restricted PDFs are rejected. It doesn't translate, make M4B audiobook files or interpret charts and equations, and complex layouts may need manual correction. The terms ask you to convert only material you own or have permission to use, and say converting a document doesn't grant any right to publish the audio.

The home page and FAQ say the workspace shows which voices and conversion options are available at the moment, and the privacy policy describes audio generation as something that runs "when generation is enabled". The voice preview on the home page also showed as pending. Check the workspace with the free allowance before buying credits.

One credit covers one character of the final text, spaces and punctuation included, so the cost depends on how much you keep after editing. New accounts get 500 free characters for 7 days. A one-time pack of 25,000 characters costs $9 and lasts 180 days, and $29 a month buys 100,000 characters per billing period, with no rollover. A voice preview uses up to 500 characters. Refunds are available within 7 days if none of that payment's allowance has been used.

Zhang Yu runs the site as an individual. The privacy policy, effective November 7, 2026, says the original PDF stays on your device, but the confirmed text goes through a service called YIR to the selected speech provider, so conversion isn't fully on-device. Saved job text and audio are available for 30 days. YIR says it deletes generated files after seven days, keeps submitted text while the submitting YIR account exists and doesn't train models on it, and deleting a job on this site doesn't remove the providers' copies. Sign-in is through Google, payments go through Waffo, and Google Analytics cookies stay off until you allow them.