Speech-to-Text Hints
Last updated
Was this helpful?
Last updated
Was this helpful?
Speech-to-Text Hints, sometimes also referred to as Phrase Lists, are instrumental in guiding the Automatic Speech Recognition (ASR) system to better comprehend words or phrases that are more likely to occur in your dialogues.
Implementing STT Hints significantly bolsters the accuracy of the ASR system, particularly when discerning phonetically similar utterances. This is invaluable for accurately capturing proper nouns such as names and locations, as well as industry-specific terminology that often poses spelling challenges. For an in-depth understanding of STT Hints and their implementation, you may refer .
To take advantage of Speech-to-Text Hints, make a service call to our dedicated Speech-to-Text Hints Service. Use the following technical specifications:
continue
: Default.
empty_stt
: Triggered when no hints are provided in the list.
If your project language is German (de-DE), take advantage of the specialized STT model called Given name (hints) that enhances further the recognition results when using STT hints. Refer to Conversation Logic, section Intents for further details on how to activate it.
Imagine a scenario in which STT hints are employed at the beginning of a sub-dialogue and are cleared as you exit that dialogue sequence. Illustrated below is a straightforward workflow exemplifying this process: