Hey folks! I'm looking for a more efficient way to keep my Supabase project updated in Claude as I build it out. I've been running into some issues because I'm constantly making changes to the database—adding columns, creating new tables, and updating RLS policies and functions.
My project's gotten pretty large, and while Claude's project system allows me to save context for relevant code generation, I'm unable to provide my entire codebase now. Thankfully, I have a modular setup and use repomix.com to create repos for the modules, which I can upload to Claude as needed. This has been going smoothly!
However, my backend workflow with Supabase isn't as seamless. I typically have to run five different SQL commands to export:
- RLS Policies
- Trigger information
- Functions
- Foreign Key Relationships
- Tables and Columns
After exporting, I give Claude these files (and sometimes I have to convert .csv files to .txt files since Claude has issues reading them) along with context from previous versions to generate updated versions for the Project Knowledge. This process takes about a week for larger updates, and it's quite tedious—sometimes Claude even misses information. Any tips on a better way to keep Claude in sync with my database?
1 Answer
You’re in luck! You can make your life way easier by using the Supabase MCP server. Here's how:
1. Install the Supabase MCP.
2. Just tell Claude to check the schema using commands like `list_organizations` and `list_tables`.
3. That’s it! You're all set!
Thanks a ton! This seems like it will really streamline my process.
I think I owe you a drink for this suggestion!