🪔 Su-śrotā — Sanskrit ASR

Transcribe Sanskrit chant and prose audio to text using prathoshap/sushrota-sanskrit-asr — an IndicConformer-based ASR model (~129M params) tuned for śāstric and recitational Sanskrit.

Upload a clip, record one from the mic, or click an example below. Audio is automatically resampled to 16 kHz mono.