Announcements
Feature: Deepgram Speech Recognition
4 min
support for this feature is in beta and should be used with caution this feature is not yet available in our enterprise environment we are pleased to announce the beta availability of deepgram as a speech recognition provider in our cloud environments this integration delivers significant improvements to the conversational experience with your digital humans key improvements improved accuracy deepgram's speech recognition model's provide more accurate recognition of what you're saying, resulting in fewer misunderstandings and a smoother conversation flow interruption support you can now naturally interrupt the digital human while they are speaking the character will stop and listen to what you have to say, creating a more human like interaction better pause handling improved handling of brief pauses during an utterance configurable utterance detection fine tune end of utterance timeouts and other settings to match your specific use case how to enable deepgram to switch to deepgram speech recognition, set speechrecognitionprovider to "deepgram" in your configuration the default settings are optimized for most use cases hosted experience window\ uneeqinteractionsoptions = { speechrecognitionprovider "deepgram" } npm package new uneeq({ speechrecognitionprovider "deepgram" }); configuration options deepgram works well with default settings, but if you need to fine tune the behavior, additional configuration options are available see https //docs uneeq io/configuration options#aqeee for details