Wow, this is amazing! As a songwriter, the only feature that I’d really need to make it an insta-buy is always-recording audio as well (and then you bookmark it and it only keeps what you sang when you started playing, since obviously audio takes way more storage space). This is probably way out of scope right now, but just adding my feedback. The always-recording piano is brilliant, but I’m not sure how useful it’d be for me without the melodies I’m improvising to go along with it. Honestly, if I wasn’t broke, I’d invest in you building that feature (or entirely separate product?) because it’d be such a gamechanger (and then you could sell it to people who play any instrument).
What kind of quality do you need from the voice recordings? Adding an integrated microphone to that enclosure would be doable but will be limited by quality and where you can place it. I'm asking because I never have a microphone while playing, but I do sing so having a rough idea what I was singing would be nice but not vital.
Seems the ESP32 can to Speex encoding so I guess Chip could integrate this as well. If he manages to earn money on this.