Fact · Languages and accents
30seconds
When no language is set, Whisper's transcribe function detects the spoken language from up to the first 30 seconds of audio.
Checked against its source on October 17, 2026
In the source's own words
“Detecting language using up to the first 30 seconds”
whisper/transcribe.py: decoding defaults and command-line options (openai/whisper), OpenAI, GitHub (openai/whisper), 2026, transcribe() function, language detection block.
Open the original- What it measures
- The code sets the language once from that window, then decodes the transcription with it. Language is not re-detected for later parts of the file.
- What it does not show
- This describes the open-source Whisper code. A recording that changes language after the first 30 seconds is not detected again by this code.
Source
Cited on
- Multilingual Dictation: Two Languages, Auto-Detect, MixingGuide
- Whisper Supported Languages and Accuracy by LanguageGuide
- Does dictation detect the language automatically?Question
- Does Whisper support Arabic?Question
- How good is Whisper at Hindi?Question
- How good is Whisper at Japanese?Question
- How good is Whisper at Korean?Question
- Why does Whisper transcribe in the wrong language?Question