I Built a 2,200-Page AI Tools Site in a Week

A week ago I decided to build an AI tool ranking site. Now it has over 2,200 pages. It updates itself every day, and the data is accurate.
Why I built it: I was already hunting for AI tools every day. Product Hunt, GitHub Trending, Hacker News — three tabs, constantly switching. I wanted one page that pulled all three together, ranked by what the community actually voted for. Something I’d use myself. Something other people could use too.
The first version took a day. Ugly, but it worked. Homepage with a few tools, vote counts, a link to learn more. I shipped it that same day. I have a rule: ship first, fix later. Things you never ship, you never finish.
The next few days were a series of decisions. Some good, some a complete waste of time.
Content volume is the foundation. If you’re building a tools directory, 100 pages isn’t enough. 1,000 might not be either. Google needs to see depth. But you can’t handwrite that many pages in a few weeks. So I spent most of my time figuring out how to produce content that is genuinely useful, scalable, and legible to search engines. That balance took the most thought.
The second decision was revenue. I spent days chasing affiliate programs — Awin, Impact, PartnerStack. All rejected. The site was too new, the traffic too small. Eventually it clicked: traffic first, revenue later. I submitted AdSense and stopped tinkering. Let the review run its course.
The third thing was the clock. I check the numbers every day — how many people came, where from, how many pages they read. Started at zero, then 15, now around 35 active users a day. It’s not impressive. But for a new site, it’s normal. Google indexing takes time, and no amount of refreshing makes it faster.
The most frustrating problems were the ones that looked small. Time zones, for example. The site is meant for a US audience, but at first the daily updates were running on Beijing time. The archive page showed the wrong date twice because I couldn’t get the US Eastern time difference right in my head. And the favicon — replacing a tiny icon took three days. Deployment issues, CDN caching, file format problems. These things aren’t coding. They’re just friction. And they eat most of your time.
Right now I’m waiting on three things: AdSense approval, more pages indexed by Google, and a real source of external traffic. None of them are things I can speed up. Everything that can be done, is done.
If I started over, I’d begin smaller and skip the affiliate detour. Google indexing is a slow variable — don’t burn energy worrying about it in the first few weeks.
A week ago I set out to build an AI tool review and ranking site. Today it has 2,200 pages, updates itself daily, tracks 98 tools, and gets about 35 active users. It’s a start. Where it is in three months, I’ll know in three months.