Backlist

Your backlist, read aloud. Proof-listened before you hear it.

Paste a chapter. An open voice model narrates it on this device, then a second model listens to every line and flags the ones that came out wrong, so you re-check a handful of lines instead of the whole chapter. Download a file mastered to audiobook spec.

1 The chapter

Try one:

Drop a .txt, .md or .docx or click to choose one

2 Voice and names

Narrator Pace

Pronunciation guide (the one you would send a human narrator: Name = how to say it)

3 Narrate and proof-listen

Narrator: Kokoro-82M (Apache 2.0). Listener: Whisper base.en (MIT). Both open-weight, both running in this tab, downloaded once: about 400 MB if your browser can use the graphics card (much faster), 170 MB if not. Your manuscript is never uploaded.