Description

Deepgram is a deep learning speech recognition platform designed to search for keywords in transcripts by sound and text. It enables users to streamline the identification of sounds and words, convert audio or transcripts into training datasets, train multiple patterns and analyze the outcomes, filter audio snippets with start and end times, identify terms or phrases within transcripts based on phonetic patterns, redact PCI from transcripts automatically, as well as reconstruct words from sounds through Hidden Markov Models and Neural Networks. Features include performance management, data labeling capability, friction point management, task automation, latency optimization, encoding management, and multi-lingual capability.