Puzzle Multilingual ASR Test
Model:
qwen3asr_small_v22_9ep_7292
/model
Settings
Live
Task
Transcribe
Translate
Languages (none = auto-detect)
English
中文
日本語
한국어
Français
Deutsch
Español
Русский
العربية
Português
Italiano
ไทย
Tiếng Việt
Indonesian
粵語
Монгол
Context / Prompt
Post-processor
off
common
voiceemr
ncc
pathology
VAD auto-segment
off
On silence after speech, the server finalizes the segment (commits the partial) automatically.
Enforce language
off
Restrict output to the selected language's script (needs exactly one language). E.g. English then never emits Korean/Arabic/Japanese.
Unfixed Chunks
Unfixed Tokens
Chunk Sec
Repetition Penalty
Input Gain (applies to mic & file, client-side)
0 dB
Reset 0 dB
Audio File (decoded → 16 kHz mono, gain applied, streamed through model)
Transcribe File
Start
Reset
Stop
Idle
-∞ dBFS
Clear history
Detected Language
—
Live Transcription
History