Zonos
Expressive open-source TTS with high-fidelity voice cloning.
About
Zonos is Zyphra's open-source text-to-speech suite, released as a v0.1 beta with two 1.6B models — a transformer and a Mamba-2 hybrid — under Apache 2.0. Trained on roughly 200,000 hours of speech, it clones a voice from 5–30 seconds of audio and offers controls for speaking rate, pitch, and emotions like happiness, anger, and surprise at 44kHz.
Zyphra positions it as matching proprietary TTS on quality while staying fully open. Practical limits: non-English languages beyond Chinese, Japanese, French, Spanish, and German are thinly represented, and some users report occasional alignment artifacts.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: Strong reception across independent outlets (TestingCatalog, theaidb, SourceForge); 200k-hour training claim and quality comparisons are Zyphra-reported; Manual review: genuinely open Apache 2.0, real voice-cloning controls, thin non-English coverage. License: Apache 2.0.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Strong reception across independent outlets (TestingCatalog, theaidb, SourceForge)
- · 200k-hour training claim and quality comparisons are Zyphra-reported
- · Manual review: genuinely open Apache 2.0, real voice-cloning controls, thin non-English coverage