Skip to content

Latest commit

 

History

History
18 lines (17 loc) · 974 Bytes

File metadata and controls

18 lines (17 loc) · 974 Bytes

Amazon Transcribe

  • Is an Automatic Speech Recognition (ASR) service
  • Input: audio; output: text
  • Offers various features that improve this feature:
    • Language customizations
    • Filters for privacy
    • Audience-appropriate language
    • Speaker identification
  • In addition we can configure custom vocabularies and configure language models according to our use-case
  • The product is pay as we use per second of transcribed audio
  • Products comes in different flavors such as Transcribe, Transcribe Medical and Transcribe Call Analytics
  • Amazon Transcribe use cases:
    • Full text indexing of audio to allow searching
    • Create meeting notes
    • Generate subtitles/captions and transcripts
    • With Call Analytics we can ingest phone calls to do analytics such as characteristics, summarization, categories and sentiment analysis
  • Can be integrated with our applications with APIs as well it integrates with other AWS ML services