Today I completed one of the core backend features of my AI Prompt Optimizer: the credit management system.
At first, I thought it would be as simple as decrementing a counter. It turned out to be much more than that.
What I implemented
Plan-based credit allocation (Free & Pro ready) Automatic daily credit reset for free users
Database-level credit deduction using PostgreSQL functions
Credit validation before processing optimization requests
Proper handling when users run out of credits
Integrated everything with Next.js and Supabase
What I learned
Building a SaaS isn't just about AI models or UI. A lot of the complexity comes from business logic.
While working on this feature, I learned more about: PostgreSQL Functions
Triggers
Foreign Keys
Row Level Security (RLS) Supabase RPC
Designing a scalable credit system
It took several iterations and a lot of debugging, but seeing everything work together was worth it.
Next up: payment integration, history improvements, and preparing for the beta launch.
Building in public, one feature at a time. 🚀