A couple of comments: for 水果, it only accepts shŭi gŭo, not shúi gŭo. I know that is technically correct, but the voice synthesis sounds it out with the sandhi. More a question of philosophy I guess, but thought it was worth mentioning.

Also, it might be nice to have an option that hides the characters, so you're forced to rely on hearing rather than just knowing that e.g. 多 is dūo.

Looking forward to seeing what else you do with it!