Media Server Overview

The Media Resource Control Protocol (MRCP) is a standard method for communicating across multiple applications. It is used primarily for telephony platforms to control resources such as automatic speech recognition (ASR) engines and text-to-speech (TTS) engines. MRCP provides a standards-based alternative to API-level integration of these resources.

For information about configuring the Speech Engine and Media Server with a third-party voice platform, see the Working with Voice Platforms article.

Note that you will not need to use MRCP if your platform has already done an API-level integration. If you are uncertain, contact your platform provider or Technical Support for more information.

To use the Speech Engine over MRCP, you must install the ASR/TTS as usual and also install the Media Server. The Media Server sits between an MRCP client (the application platform) and the ASR and TTS resources, handling MRCP communications from the client and passing requests to the speech resource servers.

MRCP comes in two versions: version one (MRCPv1) and version two (MRCPv2). The two are similar but incompatible. MRCPv1 uses RTSP as its session control protocol, while MRCPv2 uses SIP. The Media Server is designed to handle either or both protocols.

For the complete MRCPv1 specification, refer to RFC 4463. The finalized MRCPv2 specification is available as RFC 6787.

To get started working with MRCP, see Setup of MRCP API Server.

Note: For the Media Server to work properly, you must set the mrcp_server_ip value in the Media Server config file to the actual IP address of the machine running the Media Server, rather than the default loopback address. This value is used in the MRCP packets sent to the client.