Wizard-of-Oz simulation for multimodal (GUI+voice) user interfaces: The WOzOS platform

The Wizard of Oz Operating System is a Java-based platform that can be used in multimodal Wizard of Oz (WOz) experiments. It simulates a computer system that does not yet exist by means of one, two or more human operators that are hidden from the user, who is typically an external subject that doesn't know that the system is simulated. Currently, WOzOS simulates the (multimodal) combination of a graphical user interface and a voice-based dialogue system. The user can point at items on the screen, make inputs by voice, and the system (the operators) react to the input by sending back speech and screen output. To make the system 'speak', operators type in text and a text-to-speech system synthesizes the output.

WOzOS can be customized to meet various needs of user interaction studies and corpus collection exercises. Java programming skills are needed.

WOzOS was developed in 2003/2004 at MIT Media Lab Europe and served as exerpimental platform for empirical studies and corpus collection in the EU project FASiL (2002-2004). Our focus were (simulated) human- machine interactions in a multimodal environment. We collected corpora in three languages with WOzOS (English, Portuguese and Swedish) (available via ELRA).

Availability / Licensing

WOzOS is available as open source and licensed under the GNU Public License. You may freely download and modify it (and that's what you will need to do to conduct your own study with it!).

We ask that if you make changes, you do so in a generic way whenever feasible, and submit them to this project so others can use your contribution.

Documentation, background information

We describe the system and a study conducted with it in more detail in this paper: Cody M., et al., "A platform for multimodal Wizard of Oz user interaction studies".

The source code uses JavaDoc to provide documentation necessary for customization.


Click here and select the the .zip file to download the complete package.


If you'd like to give it a try, we're happy to answer any specific questions you might have. Please read the provided background information first. Then contact Brian Solon (solonb at tcd.ie) and David Reitter (dreitter at inf.ed.ac.uk).


Project leads

Michael Cody (emeritus),
Fred Cummins (Univ. Coll. Dublin) (emeritus)
David Reitter (Univ. Edinburgh)
Brian Solon (Trinity Coll. Dublin)


Mark Gargan Stephanie Richter Brian Solon Wei Zhu


We'd like to thank Sara Holm, Per Idoff and Roman Zielinksi (Capgemini Sweden), Nuno Beires, Rui Gomes and Luis Almeida (Portugual Telecom Inovacao), the Reinhard Blasig (ScanSoft) and Hans Dolfing (Vox Generation) for their help.