I got NCB to use as a backend to Greta after failing to get password reset emails working in Supabase in the first iteration of an app I was developing (despite thousands of Greta Tokens, consulting ChatGPT, Supabase help bot etc). That process was certainly much easier with NCB.

However, having worked with NCB for a couple of months now, I have some misgivings. To begin with, it doesn't support bulk data upload by importing a CSV file. To be fair, I did know that going in and thought I could develop at least one application without it, if not the other I have planned. However, it is a serious constraint. NCB's 'solution' is to use Zapier to bulk upload, which is frankly naive. It just isn't practical for routine upload of large datasets. And the suggestion made me wonder just how serious NCB are about any but the most trivial applications i.e., those with decent data sets. CSV upload is simply routine in other databases, including MySQL, the underlying DB in NCB. Hopefully, the just released Row-level Security (a great move by NCB and something I don't think MySQL does natively) , will enable CSV uploads to be implemented.

My concern deepened when I learned there is no practical backup/restore capability either, which is essential for any serious commercial application. NCB does perform backups to recover from their own system failures, but if you or your users corrupt or cause a DB failure, for whatever reason, you're screwed! There is no way to roll back your own database(s) to a previous state. NCB's suggested remedy, again naïve, is to export your Schema, tables and data. That's just not a practical daily exercise, and even if it were, there is no way to re-import your data, as NCB doesn't support the SQL INSERT command, nor as previously mentioned, can you import a CSV. So, then you're properly screwed! Again, perhaps RLS will enable future backup/recovery.

Documentation is limited as other reviewers have said. And the help bot doesn't seem to know anything e.g., 'How can I change the name of a database? ' Response: 'Can you please rephrase or clarify your question?' is just one example! The tutorial videos are helpful as far as they go, however they are hard to understand. It would help if they used a voice synthesizer to produce unaccented speech. Support is responsive to email, but I've found the advice quite cryptic for non-techos like me. I paste responses into ChatGPT and have it explain to me more fully.

I appreciate NCB is a work in progress and I'm prepared to cut it some slack because of that. However, a roadmap would useful and possibly help one keep the faith. So for me, unless the bulk data upload and backup/recovery issues are addressed I don't see myself going into production with NCB. I'll simply use it as a development environment and use ChatGPT or something to transition the schema to a more robust environment. FYI, that is how I got my original Supabase schema into NCB.