The puppet project intends to use the IBM Watson speech-to-text API to allow users to create their own text-to-speech agents.
By providing audio sample files a database of their words will be created to be used for text to speech.
Eventually this project will support generation of new words from audio samples.