I built an AI Chrome extension at 21 — spent tonsof hours on it, and nobody is using it. Roast me, help me.

I'm going to be honest in a way most "I built a thing" posts aren't: zero users. Not "early traction." Not "still growing." Zero. And it's eating at me.

I'm Daniel, 21 years old, from Honduras. I spent the last several months building Universal AI Sync — a Chrome extension that lets you sync your conversation from one AI to another in one click. Open ChatGPT and Claude side by side, click Sync, and your context transfers over so you don't have to re-explain everything from scratch. Supports ChatGPT, Claude, Gemini, and Copilot.

Extension: https://chromewebstore.google.com/detail/universal-ai-sync/hnfdhhaignnapkpkkhjocckiklgedaln

Landing page: universalaisync.com

I'm posting this to ask for honest feedback. What's wrong with it? Why wouldn't you use it? What would make you actually install this?

What I built and why
I kept switching between AI tools mid-conversation and losing all context. Every time I moved to a different AI I had to paste in background, re-explain the project, summarize what we'd already figured out. It was annoying enough that I thought — someone must have solved this. They hadn't.

So I built it. Content scripts on each AI platform that scrape the conversation, a service worker that coordinates the transfer, a backend on Railway with Supabase for the license system. Free tier, premium tier ($5/month), PayPal subscriptions.

It works. I use it myself every day. But nobody else is.

The part that actually broke me
The hardest technical part wasn't the extension architecture or the backend. It was that the DOM on every AI platform changes constantly.

ChatGPT is a Next.js app. Gemini re-renders differently in different states. Claude changed their interface structure on me twice while I was building. Every few weeks something breaks because a class name changed or a container got restructured. I'd push a build, feel good about it, and a week later the scraper would silently stop working.

I had a night around hour 200 where I was at my desk at 1am, ChatGPT had just changed their UI again, and I genuinely wanted to delete the repo. I'd already spent so much time. Every time I fixed the selectors, I knew they'd break again because I'm building on interfaces I don't control.

I didn't quit. I kept fixing it. I built more resilient fallback strategies, mutation observers, multiple selector approaches. Got better at it.

But that feeling — what if this is all pointless — never fully went away. And right now, looking at zero installs, it's louder than ever.

The payment processor wall (nobody talks about this if you're not from a developing country)
I finished the product and needed to charge money. Simple, right?

Stripe: Honduras not supported for receiving payments.
LemonSqueezy: Honduras not supported.
Paddle: Honduras not supported.

I went through basically every processor I could find. Some don't even tell you upfront — you set up the whole account, configure your product, and then when you try to enable payouts you hit the wall.

I ended up with PayPal subscriptions — not because it's good (the webhook API is annoying, the sandbox is inconsistent), but because it's the one option that actually works from here. If you're building from a country where Stripe just works, that's a privilege you probably don't think about. I think about it a lot.

The stack
Extension: Manifest V3, vanilla JS, Vite 8 (had to write a custom build.mjs because Vite 8 switched to Rolldown internally and it breaks multi-entry IIFE builds — each entry now builds in its own vite.build() call)
Content scripts: One per platform, each a self-contained IIFE bundle
Backend: Express on Railway
Database: Supabase (license keys with device binding, sync cooldowns for free tier)
Payments: PayPal
What I actually need from you
I don't know if the problem is:

The idea (does anyone actually care about this?)
The landing page (is the value prop unclear?)
Distribution (I have no audience, no Twitter following, nothing)
The product itself (is the UX confusing?)
All of the above
I've been heads-down building for months and I think I've lost perspective. I can't tell anymore if this is a product people want and I just haven't found them, or if I've built something nobody needs.

If you install it and it's broken or confusing, tell me. If the landing page makes no sense, tell me. If this idea is fundamentally flawed, tell me that too — I'd rather know now than keep going in the wrong direction.

Extension: https://chromewebstore.google.com/detail/universal-ai-sync/hnfdhhaignnapkpkkhjocckiklgedaln

Landing page: universalaisync.com

Thanks for reading.