If you plan on using MaryTTS, you need to have an MaryTTS server running. For this tutorial we will focus on using MaryTTS for speech.

• Go to http://mary.dfki.de/download/index.html and install MARY TTS v5.1.2.
• After the installation, navigate to the folder marytts-5.1.2/bin/ and run marytts-component-installer.bat. To chose the languages and the voices to install go to the <GRETA_DIR>/bin/Common/Data/characters/. Here you will find .ini files for each character where it is specified the opnemary details (see image below).

• After installed the voiced you can run the marytts-server.bat file to start the MaryTTS server. Once the server is up and running, you will be able to see the port number.

• You can change and verify the address and the port number of your server in the file <GRETA_DIR>/bin/Greta.ini. You can also check the character files in <GRETA_DIR>/bin/Common/Data/characters/ to verify which voices are needed and used.

• Now we have to add the MaryTTS Client module to the agent in the configuration window. Add the module from 'TTS' category under 'ADD' tab and connect it to the TTS module.

In order to use marytts on Linux/OSX go to https://github.com/marytts/marytts-installer , clone the repository and follow the instructions Use the following command to install the voice that GRETA uses as default voice:

marytts install voice-cmu-slt

Then run the following command to start marytts:

marytts server