Text to Speech by Language
Most browser TTS tools ship English and stop. Quick TTS ships four engines across sixteen locales — but the coverage is uneven, and which engine speaks your language is not obvious from any dropdown. These pages say plainly which engines work, which are missing, and why.
Coverage at a glance
| Language | Piper (no GPU needed) | Supertonic HD | The short version |
|---|---|---|---|
| Korean 한국어 | — | Yes — 44.1kHz | HD only — no Piper voice exists |
| Japanese 日本語 | — | Yes — 44.1kHz | HD only — no Piper voice exists |
| Chinese 中文 | zh_CN-huayan-medium |
— | Piper only — no Supertonic Chinese |
| Spanish Español | es_ES-davefx-medium +1 |
Yes — 44.1kHz | Both neural tiers |
| Russian Русский | ru_RU-irina-medium |
Yes — 44.1kHz | Both neural tiers |
| Vietnamese Tiếng Việt | vi_VN-vais1000-medium |
Yes — 44.1kHz | Both neural tiers |
Kokoro-82M is missing from that table on purpose: it synthesizes English only and is hidden on every other language, so listing it per-language would only ever mean "no".
Why the gaps run in both directions
It would be tidier if one engine were simply best everywhere. It is not. Supertonic HD is the highest-quality engine and it is the only neural option for Korean and Japanese — but it ships no Chinese at all, so Mandarin runs on Piper, the engine that is elsewhere the compromise choice. Piper covers thirteen locales and needs no GPU; Supertonic covers fifteen and needs WebGPU with no WebAssembly fallback. Neither is a superset of the other.
The practical consequence: the engine worth using depends on your language and your hardware, in that order. Each page below states both.
Pick a language
- Korean text to speech — 한국어. Supertonic HD is the only neural voice.
- Japanese text to speech — 日本語. Supertonic HD is the only neural voice.
- Chinese text to speech — 中文. Piper is the only neural voice.
- Spanish text to speech — Español. Both neural tiers available.
- Russian text to speech — Русский. Both neural tiers available.
- Vietnamese text to speech — Tiếng Việt. Both neural tiers available.
Every one of them runs in the browser with nothing uploaded, no account, and no character limit. If you want the architecture rather than the language detail, how in-browser text-to-speech works covers it, and the comparison page puts Quick TTS next to the other local-first TTS tools.