Cantari

What Cantari is

Cantari is a text-to-speech studio that routes scripts to multiple voice engines to generate audio. Its listed engines are Gemini Flash, Kokoro, Grok Voice, MAI Voice 2, and Zonos, and each generation can contain up to 30,000 characters. The service is available on the web and Windows and lists support for 24 languages and speech controls. Free accounts get MP3 export; paid plans add WAV export. Every plan includes worldwide commercial use, no watermark, and ownership of generated audio. The free plan costs nothing, billed Forever, with unlimited Kokoro, about 10 premium minutes monthly (10,000 characters), MP3 export, and Scribe for Windows. Creator is $15 a month and adds all engines, about 10 hours monthly (600,000 characters), and MP3 and WAV export. Studio is $49 a month and lists about 30 hours monthly (1,800,000 characters), audiobook studio, dubbing and translation, and both export formats. Cantari also lists speech-to-text and sound and music tools. Voice cloning is marked as coming soon, and there is no public API yet.

Who it suits

Cantari suits people generating speech audio from scripts who want multiple listed engines, speech controls, and commercial-use rights. Its Windows Scribe app may also suit people who want push-to-talk dictation that places cleaned text at the cursor.

Limits

The Free plan includes about 10 premium minutes per month, or 10,000 characters, alongside unlimited Kokoro. Creator lists about 10 hours per month and Studio about 30 hours per month; generation supports up to 30,000 characters at a time.

Getting started

1. Visit the Cantari website.
2. Choose the Free plan or a paid plan.
3. Use the web studio to select an available engine and generate audio from a script.
4. Export MP3 on Free, or MP3 and WAV on paid plans.
5. On Windows, use Cantari Scribe for push-to-talk dictation that inserts cleaned text at the cursor.

Verdict

Cantari offers script-to-audio generation through several engines, with free and paid plans and commercial-use rights. The API is not yet public, and voice cloning remains planned rather than available.