Deepgram now offers the ability to restrict which languages will be detected by Deepgram's language detection algorithm.

If you currently use the detect_language=true feature of Deepgram but your audio files are, for example, only in English and Spanish, you can improve the accuracy of your results by restricting which languages will be detected.

Here's an example of how to restrict the detectable languages to only English (en) and Spanish (es):

curl \
  --request POST \
  --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
  --header 'Content-Type: audio/wav' \
  --data-binary @youraudio.wav \
  --url 'https://api.deepgram.com/v1/listen?detect_language=en&detect_language=es'

This functionality is available for pre-recorded audio and Deepgram's models (not Whisper).

For more information, head over to our developer documentation.

Stop building work-arounds for STT systems that don't work.

Start FreeTalk to an expert