WitSound Technologies uses Automatic Speech Recognition (ASR) deep learning/ AI models as a fundamental element of our technology stack. ASR models are specifically designed to transcribe spoken language into written text.On the other hand, Large Language Models (LLMs), like OpenAI’s GPT, concentrate on generating coherent and contextually appropriate text. LLMs are tools that can be employed for various natural language understanding tasks. In contrast, ASR models are specifically tailored to cater to speech recognition applications.

WitScript supports a wide range of audio formats and is compatible with wav files with up to 512 channels. It marks discrepancies between the original script and recorded audio, streamlining the QA/QC process.

WitScript supports 32 languages: English, German, Spanish, French, Hindi, Tamil, Italian, Bulgarian, Croatian, Czech, Danish, Dutch, Estonian, Finnish, Galician, Greek, Hungarian, Indonesian, Latvian, Lithuanian, Malay, Norwegian (Bokmål), Polish, Portuguese, Romanian, Russian, Slovakian, Slovenian, Swedish, Turkish, Catalan and Vietnamese.