Inspiration

I've personally been burned by platforms that called themselves "free" and then locked everything behind a paywall. I kept seeing the same frustration in other students around me. I wanted to build the platform I wished existed: one place where you can actually trust that everything is free, curated by someone who gets why that matters.

What it does

LearnMatch is a mobile-first web app where students in grades 6–12 can get personalized recommendations for genuinely free learning tools, with no paywalls or surprises. You pick your grade, subjects, and interests in a short onboarding flow, and the app returns verified resources with a custom book rating and a "why it's for you" line so you know exactly why each tool was matched to you.

How we built it

I used Claude Code to go through a full software design process. I used it for scoping the idea, writing a product requirements doc, building a complete technical spec, and breaking it into a 13-step build checklist before writing a single line of code. The app itself is planned as vanilla HTML/CSS/JS with Tailwind via CDN, localStorage/sessionStorage for persistence, and no build tools.

Challenges we ran into

The hardest moment wasn't technical, but it was when someone challenged me that AI tools already do resource curation, making the whole project pointless. I had to genuinely work through whether what I was building had real value, and what came out of that was actually the clearest version of the idea: the trust and the human curation is the whole point, and no AI chat interface replaces that for my target user.

Accomplishments that we're proud of

I didn't write much code, but I produced a scope doc, a full PRD, a complete technical spec, and a 13-step build checklist detailed enough that any developer could pick it up and build without asking a single question. Going from a rough idea to something that thorough is something I'm genuinely proud of, and it taught me that the planning process is what separates projects that actually get built from ones that stay as ideas.

What we learned

I learned that the quality of what an AI coding agent produces is directly tied to the quality of what you give it and that means doing the hard thinking upfront, not skipping to the code. Scoping, writing user stories, defining every edge case before building: that's the process that gets real outputs.

What's next for LearnMatch

The first priority is deploying it live on Netlify and wiring up the resource request form so students can actually submit what they're looking for. From there: expanding the curated database, adding Google sign-in, and eventually letting the community vote on resource ratings so trust is built collectively, not just by me.

Built With

  • claude-code
  • claude-pluggin
Share this project:

Updates

Submission history