The Bing Speech Recognition Control enables a Windows 8, Windows 8.1, or Windows RT device to convert audio speech input to written text. It does this by receiving audio data from a microphone, sending the audio data to a web service for analysis, and then returning its best interpretations of user speech as text.
When you embed the Bing Speech Recognition Control in an app, you control how your app interprets the text returned by the Speech Recognition Service. By linking to recognized patterns, such as e-mails addresses and URIs, you can link that text to common actions such as sending an e-mail or visiting a web site. You can also combine Speech Recognition with other services such as Bing Translator or Bing Search.
The SpeechRecognizerUx control appears as a black bar when you start a speech recognition session, and contains the necessary UI to monitor or end speech recognition. You can add the control and API to any Windows Store application.
If you want to create a custom speech recognition UI, you can use the APIs in the Bing.Speech namespace without putting a SpeechRecognizerUx control on the page. Instead, you would use the methods and events of the SpeechRecognizer class with your chosen UI elements.
Developers can install the Bing Speech Recognition control for Windows 8 or for Windows 8.1 directly from the Visual Studio Gallery, or from the Extension Manager in Visual Studio. To gain access to the Bing Speech Recognition Control and the web service that it depends on, you must subscribe at The Windows Azure Data Marketplace, and then register your app. The subscription is free for the first 500,000 service calls per month.
Read More http://msdn.microsoft.com/en-us/library/dn303461.aspx