One interesting API I came across was Houndify:
This API allowed various smart-communication such as speech-recognition, natural language understanding, and audio identification. This is a handy API that can be used for a variety of tools. For example, this could be used for tools that need speech input, like Siri from Apple, and respond in a relevant and useful way. There are so many possibilities with this API.
Another API that I personally found more interesting was Musixmatch, which is an API that "permits you to search and retrieve lyrics in the simplest possible way."
As someone who karaokes in her apartment every day, I would find this API very useful. I often find myself get really frustrated when I have to frantically search up a lyrics to a good song that starts playing. The process of me searching the lyrics up just takes way too long.
With these two APIs, I can somehow create an app where it detects the audio (music in this case) and identifies the specific song, and then display the lyrics in a karoake format.
As I looked through these API info pages, I noticed that most if not all required a payment of some sort to actually have some control over the interface. Also, all them had specific copyright terms and warnings. Especially with Musixmatch, I would dealing with a lot of complex copyright rules since the database is comprised of music.