DontBuild It
AI-Powered Idea Validator
Two weeks ago I launched DontBuild.it - a tool that tells you whether to BUILD, PIVOT, or DON'T BUILD your startup idea, backed by real signals from Reddit, Product Hunt, Indie Hackers, and Hacker News.
The problem I was solving for myself
I realized I was stuck in the AI hype loop.
Every week a new idea. Every week thinking, “this one could be huge.”
But I had no real way to tell what was actually worth building.
Was it a real problem?
Would anyone pay for it?
Or was the market already crowded?
I didn’t need more opinions. I needed proof.
So I built a tool that does the research for me.
How it works
You describe your idea (the problem you're solving, your target user, how you plan to monetize)
The tool scrapes Reddit threads, PH launches, IH posts, and HN discussions in real time
An AI model analyzes the signals - buyer language, pain indicators, market saturation, willingness to pay
You get a verdict: BUILD / PIVOT / DON'T BUILD / INSUFFICIENT SIGNAL - with the actual evidence behind it
No fluff. No "great idea, here are 10 tips."
Just a clear call and reasons you can act on.
What I learned building it
The hardest part wasn't the scraping or the AI, it was calibrating the scoring. Early versions were too optimistic. Almost everything came back BUILD. I had to tune the model to be genuinely harsh, not encouraging.
Also: in-memory rate limiting resets on every deploy. Learned that the hard way.
Where it stands
Launched Friday, Feb 20
Free preview (partial results), full report for $4.99 for 3 analyses
Built in 2 weeks: React + Node + PostgreSQL + Railway + Cloudflare
Honest ask
Run your current idea through it and tell me if the verdict felt accurate. I'm especially curious about false positives, cases where it said BUILD but you already know it's a bad idea.
Comment