Deepgram is the world’s first and only Enterprise Automatic Speech Recognition (ASR) platform that uses end-to-end deep learning to transcribe audio in batch and real-time. Optimized to run on GPUs, Deepgram is the fastest, most efficient, and most accurate ASR tool available. Deepgram is an API based interface offering options for cloud and self-hosted deployments. Deepgram achieves higher accuracy for its customers by building Custom Trained Speech Models specifically trained to understand their enterprise audio.
www.deepgram.com | San Francisco, CA | Initial Investment: April 2020