How I Used AI + Google Trends to Launch 4 Micro-Sites in a Month and Built a $2,000/mo Side Income

I didn’t plan to build four websites in a month.

Honestly, I didn’t even plan to build one.

But sometime earlier this year, I fell into a rabbit hole: I was casually checking Google Trends for a separate project, and I noticed a tiny spike around a niche technical keyword. Nothing dramatic—just a small, steady curve going up. But it made me curious enough to double-click.

That tiny spike turned into the first $2,000/month I’ve ever made online consistently. And it didn’t come from one big site… but from four small, weird, AI-assisted micro-tools that I never expected to earn a dollar.

Here’s the story—messy parts included.

The Spark: When Google Trends Became My Business Partner

Google Trends isn’t glamorous. It’s not one of those dashboards with neon charts and sliders and dopamine-inducing curves. It’s almost boring—until it isn’t.

What I started doing was simple:

Type in a niche technical term
Compare it with related queries
Look for steady demand, not spikes
Ask: “Is someone clearly solving this already?”

One morning, I typed something random:“arcana calculator”

Why? Because I saw someone mention it in a forum discussing game mechanics. And to my surprise, the trend line was steady—small volume but consistent. A classic “micro-need.”

It took only one more look to realize:
People are googling it… but no one is offering a clean, modern tool.

That was site #1.

Site 1: arcanacalculator.com — My accidental warm-up project

I built the arcanacalculator.com in two evenings.
Not because I’m a programming wizard, but because AI handled 70% of the tedious parts.

What surprised me wasn’t the traffic—it was the behavior of the traffic.

People stayed.
They clicked around.
Some bookmarked the tool.

That told me everything: there’s an underserved long-tail demand for simple but high-quality micro-tools. I just needed more of them.

The Momentum: When One Idea Turns Into Four

After the first site started showing signs of life, I did another round of Google Trends and keyword exploration. I wasn’t hunting for “big opportunities.” I was looking for:

Utility
Simplicity
Problems with no pretty solution yet
Terms where people search out of frustration

This is how the next three came about—each from a random moment and a slightly obsessive curiosity.

Site 2: superscriptgenerator.net — Born from pure annoyance

One night, I wanted to type a tiny superscript “2” for a doc. I thought:“There must be a clean, instant way to generate superscripts.”

There wasn’t. At least not one that wasn’t loaded with ads and pop-ups from the early 2000s.

So I made one.

The funniest part?
This site ended up ranking the fastest. Maybe because everyone occasionally needs a superscript or subscript and nobody wants to memorize Unicode numbers.

AI Supercript generator helped generate the character sets, test device compatibility, and write some of the more complex fallback code. I just polished the UI and added examples.

Traffic grew faster than I expected. So did ad revenue.

Site 3: freevideocompressor.online — The one that unexpectedly blew up

This site was the turning point.

I noticed a constant upward curve for queries like “compress video online.” Nothing explosive—but extremely steady. Videos are getting bigger; messaging apps cap file size; creators need quick solutions.

So I asked:“Can I make a video compressor that doesn’t feel like a scam?”

Between AI-assisted WebAssembly workflows, cloud-freeencoding options, and a clean single-page UI, the video compressor tool started gaining traction. And unlike the superscript site, each user spent more time on it—which meant better ad revenue.

That’s when I realized I wasn’t building websites anymore.
I was building a portfolio of micro-utilities that quietly solve annoyances.

Site 4: deepocr.cc — Because text extraction should be simple

OCR is a crowded space, but ironically most tools fall into two categories:

Overcomplicated “enterprise” platforms
Shady-looking converters covered with intrusive ads

I wanted something that loaded fast, worked on images instantly, and didn’t make users click through seven banners or download random installers.

AI models made it dramatically easier to stand up a lightweight OCR service. I built version 1 in about a week. It’s far from perfect, but it’s clean, fast, and trustworthy. And that alone was enough to pull organic traffic.

Traffic → Money: What Actually Worked (and What Didn’t)

I monetized everything with Google AdSense at first.

And then…
I got banned.
Multiple times.

(If you’ve ever run AdSense, you probably know this pain: no explanation, no human contact, just a polite “your account is disabled.”)

So I switched to Mediavine’s “Journey” network, which has been dramatically more stable and surprisingly better-paying for my niche traffic.

That’s when the earnings leveled out to around $2,000 per month.

Not life-changing money, but enough to feel surreal considering how small each site is individually.

What AI Actually Did for Me (not the marketing version)

People online talk about AI like it magically builds income streams if you click hard enough.

That’s… not how this worked.

Here’s what AI did help with:

rapid prototyping
code snippets and debugging
generating Unicode sets
rewriting UI copy
brainstorming UX flows
validating rough technical ideas
summarizing feature gaps in competitor tools

And here’s what AI did NOT do:

validate demand
design the user experience
build brand trust
prevent ad network bans
maintain the sites
handle unpredictable user behavior

AI was a multiplier—but only because I already had a direction.

The Less Glamorous Reality

People love “overnight success” headlines. This wasn’t one.

Behind the $2,000/month:

I spent too many late nights finishing features I thought would take one hour but took six.
I rewrote the UI for my video compressor four times because it “felt off.”
I fixed obscure Safari bugs that made me question my sanity.
I got hit with abusive bot traffic.
I got banned by Google Ads more than once.

But every month, the dashboards kept ticking upward. Slowly. Quietly.

And as the traffic grew, something else grew with it:
my belief that micro-tools are still an untapped frontier.

If You're Thinking of Doing Something Similar

Don’t chase big markets.
Don’t chase viral moments.
And definitely don’t chase “AI-built business” fantasies.

Instead, look for:

Things people complain about
Tools people use but hate
Workflows people repeat 100 times a week
Tiny tasks that feel stupidly slow
Old tools that look like they haven’t been updated since 2009

Then ask the most important question:

“Can I make this 30% easier?”

If yes, you might have your next micro-site.

Closing Thoughts

I don’t know how long these sites will last. Maybe they’ll grow. Maybe they’ll die. That’s the nature of the internet.

But here’s what I learned:

AI is a power tool, not a business.
Google Trends is underrated.
Long-tail utilities are still gold.
Building several small bets beats chasing one big dream.
Simplicity earns trust faster than features.

And weirdly, the less I tried to chase money, the more the money followed.

Not a bad outcome for a month of curiosity, some late nights, and four tools that solve tiny problems surprisingly well.

搜索此博客

AnGuo