Speech Recognition Pipeline

Our Team has developed variety of tools that are in use by our customers and universities research labs. All our tools are commercially available for integration by OEM products 
   

Language Detector 

Identifies spoken Language 
 

Service providers with multi-lingual needs are challenged frequently when it comes to language detection. Our algorithm receives an audio file in any format and indicates the spoken language with high certainty  

​​

  • Supported languages are EN/ES/AR/HE

  • Available on-premise or on cloud 

  • Very fast

  • More than 93%  accuracy

Speakers Separation 

Identifies & segment two speakers   

​

Handles call center mono recording. The product is an ML Algorithm that receives a mono audio file at any format and the numbers of expected speakers and returns a timeline of who spoke when

​

  • Supported languages are EN/HE

  • Available on-premise and on cloud 

  • Very fast

  • More than 90%  accuracy