If we just could connect our Supabase project to the cursor so cursor could be aware of tables, fields, enums and anything else accessible.
Right now getting schema, and other details from Supabase to Cursr is time consuming and it could lots of issues specially regarding the field name and ENUMs
These guys tempolabs.ai has a nice supabase connect feature that does the same
This fork doesn’t have all the features of the original branch, but it can do CRUD operations and some file operations. If you have multiple projects you’ll need to run multiple mcp servers as each server is configured with the access keys and URL that belong to 1 project. Running multiple servers at the same time may confuse the agent as well so I’d be careful doing this. It’s easy enough to add/delete the server when you need it.
It requires creating a local copy of the database and then Cursor can see the data, schema, etc. locally within the IDE. You can pull and push to Supabase as needed.