RoastMyLanding is live — 8 days from idea to launch
Hey IH! 👋
I'm Ayesha (@AyeshaBuilds), a designer-developer from Bangladesh. Today I'm launching RoastMyLanding.
The backstory:
3 weeks ago I started manually roasting landing pages on Twitter and IH for free — just to validate the idea. 15+ pages roasted. Real founders gave real feedback. One IH commenter said something that changed my positioning:
"People who pay for feedback don't want a roast — they want a diagnosis."
That became the product.
What it does:
Paste any landing page URL → get a brutal 10-point conversion diagnosis in 30 seconds. Headline, CTA, social proof, pricing clarity, visual hierarchy — everything that actually kills conversions.
What I built it with:
React + Vite + Node/Express + Supabase + Groq API + DodoPayments + Vercel + Render
Pricing:
Single Roast: $5 one-time
Pro Lifetime: $29 (50 founding member spots)
Try it: https://roastmylanding.vercel.app/
PH: https://www.producthunt.com/products/roastmylanding-2?utm_source=other&utm_medium=social
Would love brutal honest feedback from the IH community — you've been the most helpful audience during this whole build.
🚀 UPDATE (April 10): A lot has changed in the last 48 hours!
After talking to some of you and looking at my conversion data, I’ve made two big pivots:
From Paid-Only to Freemium: I realized I needed to prove the AI's value first. You can now get a "Diagnosis Report" (Overall Score + Top 2 Fixes) for FREE.
Engine Upgrade: I’ve moved to a much more robust, vision-capable AI setup. While this significantly increased my cost per roast, the "brutal" accuracy and depth of the analysis are now exactly where I want them to be. Quality over margins, for now!
I’m still hunting for my 50 founding members ($29 lifetime), but I’m happy to say the "Free Tier" is already bringing in way more interesting landing pages to analyze!
Check out the new flow: https://roastmylanding.vercel.app/
FIrst off, congrats on the project and I like the idea. My main question is: Obviously this is an AI audit of the landing page. Did you contribute a bunch of case studies or examples into the LLM ? Or is the AI audit just coming straight from the Groq API you listed in your build? Essentially what I'm asking is - is there any personalized expertize being injected into the system that is creating the 10 point audit at all, it's just being completed by the LLMs base knowledge.
Great question! You're hitting on the difference between a 'wrapper' and a 'system.'
While the LLM provides the raw processing power, the audit is far from a generic 'out-of-the-box' response. Here's how the 'expert' layer is injected:
The 10-Point Framework: The AI isn't just 'guessing.' I've programmed it to evaluate pages against a specific 10-point conversion framework based on high-converting SaaS landing page benchmarks (Headline Clarity, Social Proof placement, Visual Hierarchy, etc.).
Contextual Prompting: I spent a huge amount of time on the system instructions to ensure the AI thinks like a cynical CRO expert, not a helpful assistant. It's specifically instructed to look for 'fluff' and conversion killers that standard LLMs usually overlook.
Structured Diagnosis: By moving away from a simple text-out API (like my early Groq tests) to a more robust vision-capable model, the engine now 'sees' the structure of the page, which allows it to comment on things like CTA prominence and visual flow.
So, while it uses the LLM's base knowledge of language, the methodology of the roast is the proprietary part. I'm essentially giving the AI a very specific 'CRO lens' to look through!
Thanks for such a thoughtful response. So many people are creating these products that are just spit out generic 'AI slop' and I'm still learning how to create these systems with specific instructions, framework, etc that are the key to differentiation and creating an actually valuable product and business.