§02//WRAPSRL
SUNSETTING2024–2025
WrapsRLAI-generated Rocket League car wraps.
WrapsRL started in the middle of a machine learning bootcamp and shipped shortly after I graduated. It's the project that proved to me the theory I was absorbing could turn into something people would actually pay for — in a niche I'd been deep in for years.
§01/the problem
Problem.
Custom Rocket League car wraps have always existed in the modding community, but the quality bar was high and the barrier to entry was higher. You needed to know the UV template, have real design skills, and a working modding setup. Most players just wanted a wrap that matched their vibe — without any of that.
§02/the approach
Approach.
WrapsRL treats the black-and-white UV template as a mask, then uses Gemini's region-fill to generate the decal content piece-by-piece. You upload a vibe image, pick a car, and a few seconds later you have a game-ready wrap file. Basic tier covers the common cars; Pro unlocks the full roster and higher-quality generations.
§03/what i learned
What i learned.
I was mid-bootcamp when I started this, and WrapsRL collapsed everything I was learning into one thing: generative image pipelines, prompt engineering for region-locked generation, async job queues for long-running AI calls, Stripe billing for subscription tiers, Supabase RLS for user isolation. It was my first real end-to-end SaaS and it taught me that "AI-first product" has to mean more than just "calls an API somewhere."
§stack
where it landed
Did roughly $600 in its first month, which was a real thrill. Growth stalled without a marketing engine, and the bigger story is that Rocket League is rolling out Easy Anti-Cheat — which means the modding pathway WrapsRL depends on is going away. It was a good run. The pipeline it's built on (UV-template-as-mask + region fill) is still useful for other product categories. I'm thinking about what's next.