In matters of health and personal data, not only are there strict laws on the books (e.g. HEPPA), but there remains the matter of trust.
I, too, think in terms of the "freemium" model, but I do so in the context of b-corps, the idea being that stockholders in a b-corp are not free to sue the board over financial matters. What that signals is that the entity was created with no option for exit strategies of the usual kind.
I work on open source biomedical NLP projects which aim to dive into the vast (and messy) PubMed literature, and all of the many databases on health, disease, microbiome, and etc. It seems to me that a project such as yours may end up wandering in similar directions.
I'm not well versed on business structures, so I'm having a bit of trouble following your second point about b-corps with relation to freemium.
Wow, that is exactly a path I have been thinking of taking this from the start. Vaguely, I'm interested in integrating biomedical papers linked by the conditions, therapies, and causative effects they mention and allowing discussion on them. Some of the higher-quality APIs I've found with this info do charge, so I'd be especially interested to hear about the projects you work on and if their license allows for use.
Our projects are open source under a viral license, one which does not preclude you from using them, but with care.
Edit: pinged in the feedback message.