Elmren Voice

You paste text or import TXT, Markdown, EPUB, PDF, DOCX, SRT, VTT or HTML. Scanned PDFs need a text layer first. You choose a voice, or use Script Mode to give different speakers different voices, then generate. Exports include chapter audio, M4B audiobooks with chapter navigation, a self-contained read-along HTML page per chapter that highlights words as they're spoken, and MP4 video meant for a TV or tablet.

Cloning works from a recording of roughly six to ten seconds with a matching transcript. The app checks the take for length, background noise and transcript mismatches before it accepts it. The site says to clone only your own voice or someone who has given explicit permission, such as a family member reading for a child. There are fast and quality tiers, and both run on the Mac.

According to the privacy page, imported text, recordings, cloned voices and generated audio stay in a local data folder. The network comes into play for two things: the first-launch engine and voice pack download, which runs from about 700MB to a few GB, and license activation through Payhip, with a 30-day offline grace period. Usage statistics are off unless you opt in. The website itself uses Google Fonts and PostHog analytics.

The license is $39.99 once for one Mac, with no subscription. It needs Apple silicon and macOS 13 or later. The trial allows 120 synthesis minutes within 7 days of first run, and a 7-day money-back guarantee applies after purchase.