Telephony Systems Using MRCP

Cepstral supports Media Resource Control Protocol (MRCP) version 1 and 2 giving you the ability to replace other TTS engines in any voice application platform that supports MRCP. Cepstral makes it easy to add high quality synthesized speech to your Linux based IVR, Call Center or Unified Communications system.

Chart describing the connection MRCP provides between Cepstral Text-to-Speech Software, MRCP servers, and MRCP clients

Cepstral Telephony Server

The Cepstral Telephony Server contains the Swift TTS engine, lexical preprocessor, and the user lexicon. When combined with one of our voices you have the ability to stream synthesized speech to a single call. Add components to expand the capabilities to create the perfect TTS solution for your telephony system.


Choose a Cepstral TTS voice to power your system. Purchase additional voices to offer multiple language support. Cepstral makes it easy to switch voices during a call.

Port Licenses

Stream audio to multiple, simultaneous calls by purchasing Concurrent Port licenses.

Save To File

Produce static content like IVR menu prompts and extension voicemail messages by saving the TTS audio to a WAV file. Choose from multiple frequencies and audio encodings including 8kHz u-law. Save to File is not designed for use in realtime operation during a call.

MRCP Server

Use the Cepstral plugin for UniMRCP, the world's leading MRCP platform, to connect any compliant MRCP client to the Cepstral Telephony Server.

Getting Started

Use our comprehensive tutorial to get started with the Cepstral TTS Plugin for UniMRCP.