Notetaker now accepts optional keyword hints when transcribing meetings. Use them when meetings are full of domain-specific terms that automatic transcription tends to mis-hear, such as people’s names, company names, product names, and acronyms. Passing those terms in keywords biases the recognizer toward them, and use_speaker_names_as_keywords adds known speaker names to that set automatically.
Keyword hints share the transcription_settings object with the existing language hints, but the two are independent. You can set keywords on their own, without expected_languages or fallback_language. transcription must be true for any of these settings to take effect.
-
transcription_settings.keywords— An array of up to 1000 terms to prioritize during transcription. Each term must be 1 to 200 characters and cannot contain control characters. Cannot benull. The terms bias recognition toward domain-specific vocabulary the model would otherwise mis-transcribe. -
transcription_settings.use_speaker_names_as_keywords— A boolean. Whentrue, Notetaker adds known speaker names to the keyword set so they’re transcribed accurately. Cannot benull. -
Keyword and language hints are independent —
expected_languagesis now optional. Setkeywordsalone, language hints alone, or both in the sametranscription_settingsobject.fallback_languagestill requiresexpected_languages, butexpected_languagesworks on its own. -
Available wherever
meeting_settingsis accepted — Set on individual Notetakers (POST/PATCH, grant-based or standalone), on calendars (so all matching events inherit the setting), or on individual events (which override anything inherited from the calendar). Sendtranscription_settings: nullor{}to clear inherited settings and return to default transcription behavior. -
Set transcription languages guide — Now covers keyword hints alongside language hints, including the validation limits and how the composite object is replaced as a whole.