Cracking systems and crossing rivers

Week #17 in Bangkok. I went down a rabbit hole... Three days of obsessive coding, dozens of failed attempts, and one breakthrough at 3am that changed everything. Also, I finally explored old Bangkok and fell in love with it.

The hunt: cracking Google's SERP scraping

SnappyName launched two weeks ago. Since then, 65 people have visited the site. Not bad for a free tool with zero marketing 😊

I have big plans for SnappyName. The current version is just the foundation. I want to add powerful features that will make it genuinely the best domain name generator on the web.

And to build those features, I need data. Lots of data.

The expensive, slow API problem

When indie makers need to scrape data, most turn to API providers like ScrapingBee, SerpAPI or RapidAPI. These services handle the technical complexity for you.

But they have two massive problems: they're expensive (SerpAPI charges $75/month for just 5,000 searches) and they're really slow (5-10 seconds per request).

I hate paying for expensive, slow services when I know there's a better way. Earlier this year, I built my own Python scraper to bypass Cloudflare's challenge system. Now I can scrape Cloudflare-protected domains without paying anyone. And it's way faster.

So when I realized I needed Google SERP data, I thought: "Why pay SerpAPI $75/month for slow results?”

Why nobody scrapes Google

Here's why most people don't: it's insanely hard.

Google has spent years building anti-bot systems. IP blocking, CAPTCHAs, behavioral detection, fingerprinting, rate limiting. If you try to scrape Google with a basic script, you'll get blocked in milliseconds.

That's why services like SerpAPI exist. They've cracked it, and they charge you a premium for it.

But I wanted to crack it myself.

Three days of obsession

I spent Day 1 researching approaches. Every solution was either extremely expensive (premium residential proxies cost $10+ per GB) or unreliable (proxies, even paid ones, are detected instantly today). All my tests failed.

Day 2 was dedicated to trying new configurations. I tried scraping on my own without going through intermediaries, testing different cloud providers in different countries. Nothing worked. Google's detection was catching everything. Google now blocks IP ranges by provider, and it blocks everyone: Hetzner, DigitalOcean, AWS, OHV… I've tried many of them. Even small, virtually unknown suppliers… 🤯

By midnight on Day 2, I was exhausted. But I kept thinking: "There has to be a way."

Day 3 was brutal. Sixteen hours of: “search, find, try… again” with Claude Opus 4.6 running servers everywhere. My eyes were burning. At 1am, I was still failing. Random blocks, random CAPTCHAs. Nothing stable.

At 2:30am, I was ready to give up. Maybe I should just pay for SerpAPI and move on?

Then Opus 4.6 wrote something completely innocuous about my last failures, and… Eureka! Something I hadn't considered hit me. Something completely outside the box… What we call in SEO a Black Hat technique.

Opus implemented it. I ran a test. Success.

Ran 10 more tests. All successful.

At 3am, I ran hundred consecutive requests. Zero blocks. Zero CAPTCHAs. Clean, really fast results every time.

I'd found the flaw.

The asymmetric advantage

I can't share the technical details, it’s obvious why 😄

But here's what matters: I can now scrape Google SERPs at massive scale for pennies. And it's faster than any API provider.

This is an asymmetric market advantage. While competitors pay $75-200/month for a few requests and slow results, I'm running thousands of queries for pennies with near-instant speed.

This unlocks features for SnappyName that would be economically impossible otherwise. Features requiring real-time analysis of millions of search results.

And it's not just SnappyName. My other SEO tools in development will use this same infrastructure.

Now the real building begins.

Old Bangkok: discovering the vieille ville

After days locked in my condo since the beginning of February, I needed to breathe.

My German friend Jens was back in town and suggested we explore old Bangkok. I'd been living here for months without seeing the historic center properly.

Wat Arun at sunset

We met at Sanam Chai, near Wat Pho. The vibe was immediately different: narrower streets, older buildings, temples everywhere.

We took the ferry across the Chao Phraya River to reach Wat Arun. The ferry costs 16 THB. That's €0.44. Forty cents to cross a major river in a capital city 🙃

Wat Arun at sunset was stunning. The temple's white porcelain spire glowing golden, the river reflecting everything, boats passing by. Tourist-heavy, but beautiful.

From Khao San chaos to Ram Buttri calm

We took the boat-bus back up the river and checked out Khao San Road. Yep, the famous backpacker street I'd somehow never visited.

Khao San Road is intense. Neon lights, street food vendors, bars blasting music, tourists stumbling around drunk. It's complete chaos, in the best way possible.

We walked to the neighboring street, Thanon Ram Buttri, which has a completely different energy. Quieter, more relaxed, still lively but not overwhelming.

We found a terrace, grabbed beers, and just talked about projects and life in Bangkok.

What struck me most wasn't just the temples or Khao San's chaos. It was the overall vibe. Everything felt more smooth, more organic, more human-scaled compared to modern Bangkok.

Don't get me wrong. I love Sukhumvit with its skyscrapers and massive malls. But old Bangkok has a charm the modern city can't replicate. The streets feel walkable. The pace feels slower.

I need to come back here more often.

This is week #17 in Bangkok. I cracked Google's scraping at 3am. I fell in love with old Bangkok. And SnappyName is about to get way more powerful 💪