CosyVoice 2

CosyVoice 2 Ikiganiro

Alibaba Tongyi Lab's streaming TTS reaching human-parity naturalness with near-zero latency and zero-shot cloning.

Kwiyandikisha kugirango Inyuguti

Umwandiko in Itagi: ya: Igenzura:

<speak><prosody rate="slow">Slow speech</prosody></speak>

i Byahiswemo Urugero - Kanda Kuri Gukuraho Rimwe Umwandiko:

Urugero: Bisanzwe Umwandiko, Umurongo: Itagi:. Itagi: -, Hindura Kuri Urugero: Nka Cyangwa.

Kugena (Ijambo =):

-12 +12
0.5x 2.0x
Na:,,
Audio Kugaragara. A Urugero:, Injiza Umwandiko, na Kanda.
Byaremwe
0:00
Iyimura Iyimura Ihuza in
Bwite Koresha Kuva: 5 /
TTS.ai? Abayobozi!

Ikiganiro CosyVoice 2

CosyVoice 2, from Alibaba's Tongyi Lab, was designed to make high-quality speech viable in real time. It uses a finite scalar quantization approach combined with flow matching to support streaming synthesis at extremely low latency, while reaching human-comparable naturalness that outperforms many commercial systems in subjective tests. Beyond quality, it offers zero-shot voice cloning from about 3 seconds of audio, cross-lingual synthesis, and fine-grained emotion control. Covering 8 languages with a 1,000-character cap, it's a strong fit for voice assistants, streaming TTS, and other real-time applications.

kugirango: Real-time applications, streaming TTS, voice assistants

Gushakisha byose CosyVoice 2 Amashusho

A

Mukoraporogaramu
Alibaba (Tongyi Lab)
Inyandiko y'Iyemererakoresha
Apache 2.0
Itariki
standard
Umuvuduko
medium
Guhindura izina
Oya
Ururimi:
English, Chinese, Japanese, Korean, French, German, Italian, Spanish
Inyuguti
1000

CosyVoice 2 Amashusho

Chinese Female

Chinese
Bisanzwe Female

Chinese Male

Chinese
Bisanzwe Male

English Female

English
Bisanzwe Female

English Male

English
Bisanzwe Male

French Female

French
Bisanzwe Female

German Female

German
Bisanzwe Female

Italian Female

Italian
Bisanzwe Female

Japanese Female

Japanese
Bisanzwe Female

Korean Female

Korean
Bisanzwe Female

Spanish Female

Spanish
Bisanzwe Female

CosyVoice 2 -

Yes. CosyVoice 2 uses finite scalar quantization for streaming synthesis at very low latency, which is what makes it suitable for voice assistants and real-time applications.

Yes. It offers zero-shot voice cloning from roughly 3 seconds of reference audio, plus cross-lingual synthesis and emotion control.

Yes. CosyVoice 2 is Apache 2.0 licensed. It supports 8 languages: English, Chinese, Japanese, Korean, French, German, Italian, and Spanish.
← Amashusho yose