Regarding TW + SQLite, in TiddlyWikiPharo we have been combining TW with Fossil SCM, the distributed control version system powered and used by SQLite. For this, we export, inside TiddlyWikiPharo, the tiddlers from JSON to STON , which is diff friendlier and extensible and then we save all tiddlers in Fossil. The combination is orchestrated from a data narrative written in Pharo / Lepiter (which is kind of the recently arrived equivalent of Jupyter notebooks in the Pharo world).
Combining TW, Pharo and Fossil gives us autonomous scalability and agile extensibility without going with data oligopolies.
As your project advances, we could learn how you implemented the storage backeds.
Thanks for your conceptual prototypes and keep us posted about your advances.