How adworthy.ai was built in 3 days with two skills
I get the same question every week: "Do these skills actually work, or is this just another prompt pack?"
Fair question. So here's the full breakdown of how adworthy.ai went from a half-baked idea in my notes app to a working ad creation platform with paying customers. Three days. Two skill files. No dev team.
The idea behind adworthy.ai
I kept watching small business owners waste hours in Canva trying to make decent ads. They'd scroll through templates, tweak colors for 45 minutes, export at the wrong resolution, and end up with something that looked like a college flyer. Meanwhile, the brands they were competing against had design teams cranking out polished creatives in bulk.
The gap was obvious. Build something that gives a solo founder or small marketing team access to 300+ ad templates across every major platform format, let them match styles from Pinterest URLs they already save for inspiration, and handle billing through Stripe so the whole thing runs on autopilot.
Simple enough concept. The question was whether I could actually ship it fast enough to test demand before someone else did.
The tech stack
I went with Lovable AI for the frontend and core app logic. Supabase for auth, database, and storage. Cloudflare for edge caching and DNS. And two AstroMVP skills to keep the AI agent from hallucinating its way into a mess.
The SaaS Builder skill handled the heavy lifting. Auth flows, Stripe subscription integration, dashboard layout, user management, API route structure. Every pattern in that file comes from real apps we shipped at AstroMVP over the past year. Finix, Tradezo, ReplyQ, and 8 others. So when the agent built the billing logic, it wasn't guessing. It was following the same architecture that already processes payments in production.
The Landing Page Builder skill took care of the marketing site. Above-the-fold hero, feature sections, pricing table, social proof, FAQ, footer with proper CTA placement. I gave the agent a rough wireframe description and the skill filled in the conversion patterns. The landing page was live before the end of day one.
Day-by-day breakdown
Day 1: Core app + landing page. I dropped both skill files into my Lovable AI project and described what adworthy.ai should do. The agent scaffolded the full app structure in about 20 minutes. Auth with Supabase, three subscription tiers in Stripe ($19/month starter, $49/month pro, $99/month agency), a template gallery UI, and the landing page. I spent the rest of the day refining copy and fixing small layout issues. Nothing structural.
Day 2: Template engine + Pinterest matching. This was the hardest part. The template system needed to support 300+ designs across Instagram, Facebook, Google Display, Pinterest, and LinkedIn ad formats. I built a metadata schema for each template, category tags, and a search/filter system. The Pinterest URL matching feature scrapes color palettes and layout patterns from a URL you paste in and suggests the closest templates. Took most of the day to get the matching accurate, but the SaaS Builder skill had already set up the API routes and database tables I needed, so I was only focused on the actual product logic.
Day 3: Billing, polish, launch. Stripe webhooks, subscription management, usage limits per tier, and a settings page. The SaaS Builder skill had already wired up 90% of this on day one. I just needed to customize the tier limits and add the upgrade/downgrade flow. Spent the afternoon on mobile responsiveness and ran through the whole flow twice. Pushed to production on Cloudflare around 6pm.
First paying customer signed up that evening. $19/month starter plan.
What the skills actually did vs. what I did
I want to be specific here because "built in 3 days" can sound like marketing fluff.
The skills handled: project scaffolding, auth implementation, Stripe integration patterns, database schema design, API route structure, landing page layout and conversion patterns, component architecture, and responsive design foundations.
I handled: the template engine logic, Pinterest URL scraping and matching algorithm, template metadata schema, copy and brand voice, final QA, deployment config, and Stripe account setup.
The skills didn't write every line of code. They gave the AI agent a proven blueprint so it wasn't making architectural decisions from scratch. That's the difference between getting a working app in 3 days and spending 3 weeks debugging auth flows and payment webhooks.
Numbers after the first month
- 47 registered users
- 12 paying subscribers across all three tiers
- $638 MRR
- 2,400+ ads created on the platform
- Zero server incidents (Supabase and Cloudflare handle scale without drama)
- Total cost to build: $58 in skills, $25/month for Supabase Pro, $0 for Cloudflare free tier
Not life-changing revenue. But for a product that took 3 days and under $100 to build, the unit economics work. And the growth curve is pointing up.
What I'd do differently
Honestly, not much. If I rebuilt adworthy.ai tomorrow, I'd add the shadcn Dashboard skill for the admin panel. The default dashboard the SaaS Builder produces is solid, but the shadcn skill makes data tables and charts noticeably cleaner. That's a cosmetic upgrade though, not a structural one.
I'd also start with the SEO Optimizer skill from day one instead of bolting on meta tags after launch. Lost about two weeks of indexing time because I was lazy about structured data.
Why this matters for other founders
The point of this case study isn't "look how fast I am." It's that the bottleneck for shipping software has completely shifted. The hard part used to be writing code. Now the hard part is knowing what patterns to use and making product decisions.
A skill file solves the patterns problem. It gives your AI agent the same architecture knowledge that used to live in a senior developer's head. You still need to make the product decisions, write your own copy, and do the work of understanding your customers. No skill file replaces that.
But if you've got a clear idea and you're burning time on implementation details, the SaaS Builder and Landing Page Builder together cost $58 and saved me at least 3 weeks of development time. That math works for pretty much anyone.
If you want the full context on why we shifted AstroMVP from an agency to a skill marketplace, read this post. It explains the thinking behind the whole pivot.
Frequently asked questions
Do I need to know how to code to use these skills?
No. Several adworthy.ai features were built by giving plain English descriptions to the AI agent with the skill file loaded. You need to understand what you want to build, but you don't need to write code yourself.
What AI tools work with AstroMVP skills?
Claude Code, Cursor, Windsurf, Lovable AI, and any coding agent that accepts instruction files. The skill files are plain text, so they work everywhere.
How is this different from a boilerplate or starter template?
A boilerplate gives you a static codebase you have to modify. A skill file teaches your AI agent how to build the right architecture from scratch, customized to your specific product. You end up with cleaner code because nothing is leftover scaffolding from someone else's project.
What does adworthy.ai pricing look like?
Adworthy.ai starts at $19/month for the starter plan. Pro is $49/month and agency is $99/month. Each tier increases template access and monthly ad creation limits.
Can I use the same skills to build something completely different?
Yes. The SaaS Builder skill has been used to build adworthy.ai, admix.software, and several other products that have nothing in common except the underlying architecture patterns. The skill teaches structure, not a specific product.
How long do the skills take to set up?
Under 3 minutes. Buy the skill, drop the file into your AI agent's context, and start describing what you want to build. There's no installation or configuration.