Building an MVP and trying to avoid spending weeks on backend setup
retained text
I'm working on a small AI photo app where users can upload pictures, generate different versions of them and save their favorites to their own gallery. Still pretty early but I've got the basic login flow working and I'm starting to build out the part where users can actually store and manage their images. One thing I'm trying to figure out before going much further is how to handle permissions properly. Obviously users should only be able to see their own uploads and generated images but I'm not sure where most of that logic should live. I've been looking into row level security in Postgres and from what I understand I could handle a lot of this directly in the database. But I've also seen people recommend keeping permission checks in the backend, especially once an app starts getting more complicated. For anyone building apps with user generated content are you relying mostly on RLS policies to keep data separated or are you checking permissions in your backend too? Trying to set this up properly now rather than having to rethink everything once people start using it.
document fetched 2026-11-03 00:28 · canonical https://reddit.com/r/SaaS/comments/1xm9olt/building_an_mvp_and_trying_to_avoid_spending