Sign in
Audio Service requires an API key
Sign in
Audio Service
signed in
Sign out
checking…
CosyVoice3 TTS + Whisper STT · API-compatible with audio-service (v2 engine)
Text → Speech
Speech → Text
你好！呢度係測試，今日天氣幾好。
粤语
普通话
English
长句
🎲 Random
Voice
Speaker reference
Auto-detect (yue/mandarin/en)
Cantonese (speaker_zh_yue)
Mandarin (speaker_zh_mandarin)
English (speaker_en)
Default Chinese (speaker01)
Speed
1.0
Advanced: instruct control (natural language)
Happy
Sad
Fast
LLM rewrite (numbers → words, dialect phrasing — slower)
Synthesize
Download WAV
Drop audio file here or click to choose (wav / mp3 / m4a / flac…)
Whisper model
tiny (fastest)
base
small
medium (recommended)
large (slowest, most accurate)
Smart split (long recordings)
Transcribe
Copy