Listen to Your Documents as One Natural MP3

One AI Creation Credit per 10,000 words, rounded up, with a one-credit minimum. A 6,000-word article costs one credit; a 24,000-word report costs three. You are charged when the narration starts, and refunded automatically if it fails.

Up to 30,000 words in a single narration, which is roughly three hours of audio. For anything longer, narrow the page range or split it into two narrations — you can select a page range on any PDF before you start.

PDFs, Word documents, PowerPoint decks, plain text and Markdown files, spreadsheets, saved web pages, and anything already sitting in your Scholarly library. You can also paste a URL and narrate the page behind it.

No. Text-to-Speech reads the source as written, in order, with nothing added or removed. If you want a shorter spoken version that explains the material rather than reciting it, use the audio summary generator instead.

Yes. The narration is saved to your library as a normal MP3 file, and you can download it and play it in any podcast app, media player, or car stereo. It works offline once downloaded.

A short article is usually ready in a minute or two. A full-length report can take several minutes because the audio is synthesized in pieces and then assembled into one file. It runs in the background, so you can close the tab and come back.

Yes. Over 25 voices are available, each with a distinct character and pace, and you can preview them before choosing. You can also set the playback speed between 0.75× and 1.5× before the narration is generated.

It reads the text it can extract from your document, and strips formatting characters that would otherwise be spoken aloud. Dense tables and mathematical notation do not narrate well by nature — for material like that, a study guide or audio summary is usually the better format.