From Utility Script to Enterprise-Grade Domain Verifier in 2 Days

·3 min read·...·Updated: July 7, 2025
From Utility Script to Enterprise-Grade Domain Verifier in 2 Days

"I just needed a few domains. Then I blacked out and built a system."

🛰️ The Spark

Two days ago, I needed to fetch verified domains for a list of companies as part of my day job. Nothing glamorous—just basic ops.

I figured: "Write a quick script, call it a day."

But then the questions started stacking:

  • What if the API rate-limits?
  • What if I need to retry failures intelligently?
  • What if I want this to scale?
  • What if… I modularize it?

And that's when the rabbit hole opened.


🛠️ What I Built

✅ A Full Enterprise-Grade Domain Verification Pipeline

Built with:

  • Multi-threaded parallel processing
  • Rate-limiting using a token bucket algorithm
  • Exponential backoff + jitter
  • Thread-safe persistent caching
  • Text similarity validation using RapidFuzz
  • Structured logging and real-time performance metrics
  • Disk-backed cache with Polars for blazing-fast I/O
  • Docker + Compose with full dev/prod flows
  • 90+ tests with pre-commit checks and CI config
  • CLI interface, containerized and modular

🚀 Performance gain?
From 0.1082s per match → 0.0030s per match
That's a 35.8x speedup.


⚙️ Why Crunchbase?

I used the Crunchbase API via RapidAPI because:

  • It was free.
  • I assumed the data was fresh.

Turns out? It wasn't always. So I built a validation module with text similarity scoring to check against expected results. It doesn't just fetch data—it judges it.


💡 From "Just for Me" to "This Might Be Useful"

What started as a one-off script became something worth sharing. I realized:

  • Others need this.
  • It could be generalized into a B2B verification layer.
  • It might deserve a GUI + billing layer.

"Would you use this if it had Stripe billing + a FastAPI GUI?"
(If you're in sales ops, lead gen, fraud prevention, or data hygiene—this might save you hours.)


🧱 Tech Stack

  • Python 3.9+
  • RapidFuzz + Polars + DiskCache + orjson
  • Docker + Docker Compose
  • pytest + coverage + pre-commit
  • CLI-ready, containerized, modular

🧬 What's Next?

Coming soon:

  • 🧠 LLM-based domain validation and judgment scoring
  • 🌐 GUI via FastAPI
  • 💸 Stripe integration for self-serve billing
  • 📦 SaaS packaging for teams

📬 Want Early Access?

I'm building this in public.

Drop your email via contact form or DM me @mebroskee if:

  • You want to try it out early
  • You work with messy domain data
  • You want to help shape the roadmap

🔗 GitHub + README

GitHub

Fork it, remix it, deploy it. It's yours now.


⚡ Closing Reflection

I didn't know how to build like this three months ago.
I've been learning in the open, guided by curiosity, LLMs, and the will to just build what I need.

If you're somewhere between "I have an idea" and "I ship systems now"—then welcome. Let's build.

Luiz Frias

Luiz Frias

AI architect and systems thinking practitioner with deep experience in MLOps and organizational AI transformation.

Comments