Hey folks! I'm looking for a better way to keep my Supabase database in sync with Claude. As my project grows, I'm constantly updating my database—adding new columns, new tables, setting up RLS policies, and creating functions. Claude has a "projects" system that helps me maintain context, but with the complexity of my codebase, I can't share everything at once. I've been using repomix.com to modularize my project and upload relevant parts to Claude, which is efficient for the front end.
However, for my backend, the process is painstaking. I have to run five SQL commands in Supabase to export RLS policies, triggers, functions, foreign key relationships, and table details. I then hand these files to Claude, which sometimes struggles with formats like .csv, forcing me to convert them to .txt. This is a recurring task, usually after big changes, and it eats into my Claude usage. I'm running into issues with Claude missing information sometimes, so any tips on streamlining this process would be much appreciated!
2 Answers
I’ve got a game changer for you! Try using the Supabase MCP server. Just install it and tell Claude to check the schema using commands like `list_organizations` and `list_tables`. It really simplifies the whole process!
Thank you! I feel that this is going to make such a difference.
Have you considered switching to Prisma? I just give Claude my Prisma schema in my prompts and it works perfectly. You might find it much easier!
Ooh, okay I think you're on to something here. Thank you, I'll check it out.
I think I owe you the adult beverage of your choice.