This paper aims at analysing the current state of knowledge regarding implementation of the multi-agent paradigm into the smart houses. It explores the architectures of both individual agents and whole multi-agent systems. The following chapters summarise chosen interesting ideas in this area.