Natural Language Processing (NLP) deals with deciphering spoken language grammar and our understanding of the same. So far my app had two or three words commands. And coincidentally they followed a pattern. Like "play movie batman" or "switch lights on". I broke it down as:

1. The device that should handle the command
2. The function
3. Optional additional data required to perform the function

However, for mood light, this pattern breaks down. One is more likely to say "color device red" or "color all green". Got to spend some time thinking of a sensible algorithm to handle this possibly using NLP.