Sentigral Transcription is an English-Language Transcription model based on the Kaldi Speech Recognition Toolkit and the Librispeech dataset.Data charges equate to approximately $0.01 per hour of audio, based on 32kHz mono WAV files and our charge of $0.04 per GB of received data. You can send any type of WAV file with any sample rate and bit depth. They'll be converted to 32kHz for our processing, and we'll send you the JSON.