Building AI-Tools: Why I Spent 3 Months Cataloging 125+ AI Tools and What I Actually Learned

Honestly, I didn't expect this project to consume three months of my life.

It started as a simple side project: I was tired of seeing the same AI tools recommended over and over again on Twitter. Every "Best AI Tools" list looked identical, featuring the same 10-15 big players that everyone already knows about. What about all the smaller, niche tools? What about the tools that solve specific problems instead of trying to be everything to everyone?

So I decided to do something crazy: catalog and test 125+ AI tools myself, one by one, and write honest reviews about what works, what doesn't, and who actually needs each tool.

Three months later, I've got some thoughts. And spoiler alert: most of what you read about "curating the best AI tools" is wrong.

How This Started

I've been building AI-powered tools for years, but I still found myself overwhelmed by the pace of new releases. Every week there's a new "AI killer app" that's supposed to change everything. But when you actually try them, most of them solve problems you don't have, or they're just wrappers around OpenAI anyway.

I wanted to create something different:

• A curated list where every tool gets an honest review
• Categorized by actual use cases, not vague buzzwords
• Pros and Cons for every entry - yes, even the popular ones have flaws
• Updated regularly as new tools emerge and old ones die

The repository is here if you want to check it out: https://github.com/kevinten10/AI-Tools

The Cataloging Process: How I Did It

I didn't just scrape Awesome-AI lists and call it a day. I actually signed up for almost every tool and tried it myself. Here's the basic workflow:

// Basic structure for each tool entry
const tool = {
  name: "Tool Name",
  description: "What it actually does (in plain English)",
  category: ["Productivity", "Code Generation"],
  pricing: {
    free: true,
    paidTier: true,
    startingPrice: "$10/month"
  },
  pros: [
    "Actually solves a specific problem",
    "Clean UI, no bloat"
  ],
  cons: [
    "Limited free tier",
    "Exporting is clunky"
  ],
  whoIsItFor: "Developers who need quick prototyping",
  myRating: 4 // 1-5
};

The whole thing is a simple Node.js project that generates a clean markdown catalog. I even added GitHub Actions to automatically update the README with the latest count.

Here's the automatic update script I use:

const fs = require('fs');
const path = require('path');

// Read all tool entries
const toolsDir = path.join(__dirname, 'tools');
const toolFiles = fs.readdirSync(toolsDir);
let totalTools = 0;

toolFiles.forEach(file => {
  if (file.endsWith('.json')) {
    const tools = JSON.parse(fs.readFileSync(path.join(toolsDir, file), 'utf8'));
    totalTools += tools.length;
  }
});

// Update README with current count
let readme = fs.readFileSync(path.join(__dirname, '../README.md'), 'utf8');
readme = readme.replace(
  /## Currently Cataloging: \d+ AI Tools/,
  `## Currently Cataloging: ${totalTools} AI Tools`
);

fs.writeFileSync(path.join(__dirname, '../README.md'), readme);
console.log(`Updated README: ${totalTools} tools total`);

Simple, effective, and it keeps everything maintainable. No fancy database needed - just JSON files that are easy to edit.

What I Learned (The Surprising Stuff)

Okay, here's the real value. After testing 125+ tools, these are the lessons that stuck with me.

1. Most AI Tools Are Just "Hanging Fruit"

Here's the dirty secret: most AI tools solve problems that weren't really problems to begin with.

Think about it: do you really need an AI to write your LinkedIn headlines for you? Is that actually slowing you down, or is it just a neat trick that someone monetized?

I'd estimate that about 60% of the AI tools I tested fall into this category: they automate a tiny step that wasn't worth automating. It's not that they're bad - it's just that you'll probably never use them after the first week.

Pros: They're easy to build, easy to market, and they make great Twitter ads.
Cons: They don't provide enough recurring value to keep you subscribed.

2. You Only Actually Need About 5-10 AI Tools

I'm not kidding. After testing over a hundred tools, I still use basically the same set I started with:

• ChatGPT / Claude for general reasoning
• GitHub Copilot for coding
• MidJourney for images when I need them
• Notion AI for notes整理
• Maybe one or two niche tools for specific work things

That's it. The rest either do the same thing better, or they solve problems I don't have.

Does that mean the other 115+ tools are useless? No. But they're useless to me. And that's okay. The right tool depends on your specific needs.

3. Price Doesn't Equal Quality

I tested tools that cost $50/month that were basically just fancy prompts wrapped in a UI. I also tested completely free tools that were more useful than some paid alternatives.

For example:

• Free tool that's amazing: [Placeholder - I'll keep this updated in the repo] - solves a specific problem perfectly, costs nothing
• Expensive tool that disappointed: Another popular one that shall remain nameless - $30/month, but the AI misses the mark more often than not

My rule of thumb now: if an AI tool costs more than $20/month, it better save me at least an hour a week. Otherwise, it's not worth it.

4. Maintaining This Is More Work Than I Expected

I thought "oh, I'll just add a few tools every week and it'll be fine." Nope.

Things change:

• Tools get acquired
• Pricing models change
• Features get removed (or added)
• Entire services shut down

Maintaining an accurate, up-to-date list is actually a full-time job in itself. I started this as a side project, and it's become something that takes me 4-5 hours a week just to keep current.

Is it worth it? For me, yes - I learn something new every time. But if you're thinking about doing something similar, go in with your eyes open.

5. Discovery Is Still Hard

Even with 125+ tools categorized, I still find out about new tools every week. The AI space moves that fast.

What this project has taught me is that curations are inherently incomplete. By the time you publish a "Best Of" list, it's already outdated.

The value isn't in being comprehensive - it's in being honest. A smaller list with real, tested reviews is worth more than a huge list with generic descriptions copied from the tool's website.

Pros and Cons of This Whole Endeavor

Let me be honest - this project hasn't been all sunshine and roses.

What Worked Well

✅ I actually discovered some hidden gems that I probably would have never found otherwise. There are some really smart people building really useful tools that just don't have big marketing budgets.

✅ It forced me to try things outside my comfort zone. I don't typically use AI for content creation or image generation, but testing these tools expanded how I think about AI augmentation.

✅ The community feedback has been incredible. People actually use this list! I've gotten dozens of messages from developers and founders saying it helped them find a tool that solved their problem. That's incredibly rewarding.

✅ Simple architecture wins. Keeping everything as JSON files in Git means version control is free, collaboration is easy, and I don't have to worry about hosting or databases.

What Didn't Work So Well

❌ It took way longer than I expected. Three months. Three. Months. I thought this would take a month. Nope.

❌ Some tools wouldn't let me sign up. Waitlisted, required LinkedIn requests, wanted me to "book a demo" - look, if I'm just testing your tool to review it, I'm not booking a 30-minute demo. That's an automatic no from me.

❌ Keeping everything updated is draining. Like I said earlier, this is ongoing work. Some weeks I don't have the energy to add new tools.

❌ Discoverability within the list is still hard. Even with categories, finding what you need when you need it isn't as easy as I'd like. I'm still working on better organization.

Who Should Actually Use This List?

If you're building AI tools yourself, this list is great for:

• Competitor research - see what else is out there
• Inspiration - maybe you'll spot a gap you can fill
• Understanding what problems people are actually solving with AI

If you're a developer looking for AI tools to help you work better:

• Great starting point - I've done the initial filtering for you
• Read the pros and cons - I call out when a tool isn't what it claims to be
• Check the "who is it for" section - saves you time signing up for things that aren't for you

If you're a product manager:

• See how different tools approach UX and monetization
• Lots of good examples to study

The Surprising Conclusion

After three months and 125+ tools... I still think the best AI strategy is to master a few tools really well rather than trying every new thing that comes out.

The AI gold rush is real, but most of the gold is already claimed. The new tools that stand out are the ones that solve specific, painful problems for specific people - not the ones that try to be "the everything AI platform."

Honestly, I'm glad I did this. I learned a ton, I found some great tools I still use, and I've helped other people find tools that work for them. Would I do it again knowing how much work it would be? Maybe not. But I don't regret it.

Check It Out

The repository is here: https://github.com/kevinten10/AI-Tools

• 125+ AI tools tested and reviewed
• Categorized by use case
• Honest pros and cons for every entry
• Open source, pull requests welcome for new tools

My Question For You

I'm curious - how many AI tools do you actually pay for right now? And what's one AI tool you've found that's actually worth the money that not enough people know about?

Drop a comment below - I'm always looking for new tools to add to the list!