Interesting! You should consider making the email field optional: it’s unclear why it’s required, and maybe folks will be more comfortable sharing if it is optional.
Yes, the underlying dataset very much conditions the quality of the responses. Additionally, the retrieval strategy is also a really important factor (and that is something which I haven't had time to extensively optimize).
I'm writing a blog post that will answer your questions! Will post it here when it's fully baked.
Thanks for flagging! Actively debugging, looks like one of my DB vendors is having some uptime issues. If you retry a couple times you should (eventually) get lucky.
Hey! Not open source, largely because I don't have time to make it good enough for me to feel comfortable sharing what's otherwise pretty scrappy code, but I'm planning a blog post detailing how it was built.