Custom Speech Service


Customize Microsoft's speech-to-text engine for your application.

Create custom language models

Customize the language model of the speech recognizer by tailoring it to the vocabulary of your application and the speaking style of your users.

Create custom acoustic models

Customize the acoustic model of the speech recognizer to better match the expected environment and user population of your application.

Deploy your custom models to a speech recognition endpoint

Deploy your models to create a speech recognition endpoint that's customized to your application.

Access your endpoint from any device

Send requests to your custom endpoint using the Microsoft Cognitive Services speech client library.