I created a STT interaction using the p5 speech library. I wanted to create something kind of like a storybook or matching game that kids play, except it works for anything. I ended up combining the p5 speech recognition with the Flickr API. So when you talk to into the computer's microphone, it will be listening, and the search Flickr for an image of what you said. I think this could be quite fun for parents and kids. It is a fun interaction that puts a picture to words and helps visualize something verbal.