Cantonese · Chinese characters
Transcribe Cantonese audio to text
Cantonese (粵語) is spoken mainly in Hong Kong SAR China, Macao SAR China, and China and written in Chinese characters. Upload a recording — lecture, interview, voice message or meeting — and get a searchable transcript with timestamps in a few minutes.
Transcribe Cantonese nowHow well Cantonese works on Sprechverlauf
Cantonese works in the free trial: three transcriptions of up to 10 minutes per calendar month, no account, as text with timestamps. It runs on Whisper Large v3 Turbo, a model trained on roughly 100 languages.
With a day pass, credit or Pro, Microsoft MAI-Transcribe-2 takes over. It separates speakers, so the transcript shows who said what — useful for interviews, meetings and conversations in Cantonese. Microsoft reports an average word error rate of 5.2% across the model's 60 languages (FLEURS benchmark); individual languages score above or below that.
What matters for Cantonese
Cantonese is a tonal language: the same syllable changes meaning with pitch. That makes a clean recording especially important — reverb, background music or heavy compression blur exactly the pitch contours the meaning depends on.
Written Cantonese does not put spaces between words. The model decides where words begin and end, so it is worth checking technical terms and names in the transcript.
Cantonese is close to Chinese. On short clips, automatic detection can mix them up — select Cantonese explicitly before you start.
Tips for better Cantonese transcripts
- In step two, pick "Cantonese" instead of automatic detection if the recording is short or starts with music.
- Record close to the microphone; a clip-on mic beats a built-in laptop mic by a wide margin.
- Overlapping speech hurts every recognizer. Short pauses between speakers help more than any editing afterwards.
- Check names, numbers and technical terms: click a timestamp to hear that exact passage.
Frequently asked questions
Can I transcribe Cantonese for free?+
Yes. The free trial transcribes Cantonese without an account: three files of up to 10 minutes per calendar month, as text with timestamps. Longer files and speaker separation start at €2.99, no subscription.
Does Sprechverlauf separate speakers in Cantonese?+
Yes, with a plan. Microsoft MAI-Transcribe-2 separates speakers in Cantonese too; the transcript then shows "Speaker 1", "Speaker 2" and so on. The free trial gives plain text.
Do I have to select Cantonese before starting?+
No, the language is detected automatically. Selecting it helps with very short recordings and because Cantonese is easily confused with Chinese.
Which files can I upload?+
MP3, WAV, M4A, OGG/Opus, FLAC, WebM, AAC, MP4, MOV — upload up to 100 MB or paste a direct link to the audio file. Download the result as Word, PDF, TXT, SRT or VTT.