I spent the weekend building PageScore — an AI-powered landing page analyzer. You paste a URL, it gives you a score out of 100 and 3 specific things to fix.
How it works
- Paste any URL
- AI fetches and analyzes the actual HTML
- You get a score (0-100) + 3 actionable tips in ~30 seconds
No signup required for the free scan.
I tested it on popular sites
Here's what I found:
| Site | Score |
|---|---|
| Stripe | 75/100 |
| Vercel | 75/100 |
| Notion | 65/100 |
| Linear | 65/100 |
| Superhuman | 65/100 |
| Cal.com | 65/100 |
Even billion-dollar companies have room to improve their landing pages.
The tech stack
- Next.js 16 + Tailwind CSS 4
- GPT-4o-mini via OpenRouter for the free scan
- GPT-4o for the paid deep-dive report
- Lemon Squeezy for payments ($7 one-time)
- Resend for email delivery
- Vercel for hosting
- PostHog for analytics
The free scan costs about $0.001 per analysis. The paid report costs ~$0.03. Margins are essentially 99%.
Try it
https://pagescore-tau.vercel.app
Paste your site's URL and see what score you get. I'd love to hear:
- What score did you get?
- Was the feedback useful?
- What would make this more valuable?
Built the whole thing in a day as an experiment in validating micro-SaaS ideas quickly. Happy to answer any questions about the build process!
Top comments (0)