A work in progress nearing completion, to be shown at Rendr on May 2, 2018.
Currently, it consists of 5 stations, each of which includes a Raspberry Pi, USB speaker, USB microphone, and display screen. Each station records and passes messages to the other units, playing a kind of game of telephone. They form a modular apparatus — each unit is independent and non-networked, and the number of devices in the chain is variable. It makes use of Carnegie Mellon’s PocketSphinx speech recognition engine, the SVOX Pico text-to-speech engine, and the Python programming language to keep each unit operating on a schedule. I am constantly remaking it right now, but I’ll share video and programming documentation here after the Rendr show.