Four steps. One Mac.

Start with the words you have

Paste text straight in, drop in a file — TXT, MD, EPUB, PDF, DOCX, SRT, VTT, or HTML — or bring a character-based script for multi-speaker productions. The app keeps chapters and structure intact, and you can merge, split, or skip chapters before anything is rendered.

FORMAT NOTE — this version reads TXT, MD, EPUB, PDF, DOCX, SRT, VTT, and HTML — or just paste the text in. Scanned PDFs need a text layer to import; structure editing works the same either way.

Pick a voice — or make one familiar

The library holds 63 built-in voices — 54 Kokoro presets plus 9 Qwen voices. Or clone your own voice — or a family member's with permission — from one short, clean recording — six to ten seconds — and reuse it in every project. In Script Mode the app auto-suggests a voice for each speaker — you confirm, and can change any of them by hand.

RECORDING TIP — a quiet room, a clean start at a sentence boundary, and a transcript that matches every word. That is the whole recipe; better input beats longer input.

Your Mac becomes the engine

Speech is synthesized locally — first launch downloads the on-device packs (engine + voices, from ~700MB up to a few GB with all voice packs, once), then the studio runs fully offline. Long jobs stay visible and interruptible, and revising a paragraph lets you re-render without starting over. Read the privacy page for exactly what touches the network (almost nothing).

Open it anywhere

Keep chapter audio — including M4B chapter audiobooks — for any player, ship a self-contained HTML page with word-level highlighting for follow-along reading, render a video for the big screen, or send the finished work to Email, Telegram, or Discord (Email 25MB · Telegram 50MB · Discord 8MB). A good first project is a family audiobook — one chapter is enough to hear the whole pipeline.