Sarvam AI icon

Sarvam AI

Indian language AI APIs — transcribe, speak, and chat across 22+ Indian languages

Actions3

Speech → Speech to Text

AI-generated

Summary

Transcribe audio to text with selectable output modes

Inputs

  • binaryPropertyName (required) — Name of the binary property containing the audio file
  • sttMode (required) — Mode of speech-to-text operation (e.g., transcribe, translate, transliterate, verbatim, codemix)
  • speechToTextOptions — Configuration options for the transcription, including language code

Output shape

a single record containing the transcription result

The result includes a 'text' field with the transcribed text; additional metadata may be present depending on the selected mode.

Examples

Example 1: Transcribe a Hindi audio file and translate it to English

binaryPropertyName: data; sttMode: translate; speechToTextOptions.language_code: hi-IN

Example 2: Produce a word‑for‑word transcription of a Tamil interview

binaryPropertyName: audio; sttMode: verbatim; speechToTextOptions.language_code: ta-IN

Links

Discussion