How I Built a Website for $2 a Year Using AI Tools (No Coding)

A year ago, building a website meant hiring a developer or spending months learning to code. I did neither. I'm a banker in Inner Mongolia, China—someone who spends his days with spreadsheets, not scripts. But today, I run a fully functional, monetizable website that I built myself using AI tools. The total cost? My domain costs $1.88 per year. Hosting is free. I spent around $2 on AI credits during the building phase. That's under $4 for the first year. And if you'd told me that a year ago, I'd have said it was impossible.

The site is ainocode.top—a resource hub for non-programmers like me who want to use AI to create things. It's my experiment, my learning journal, and now my side hustle in progress. This is the story of how I built it, the mistakes I made, and why you can do it too.

Why I Decided to Build This

I've been working in banking for years. Every day I see how AI is reshaping finance—fraud detection, customer service chatbots, automated reporting—and I felt that if I didn't start learning now, I'd be left behind. But I wasn't a coder, and I didn't have months to learn. I needed a practical project that would force me to use AI in a real way, and I wanted to document my journey so others could follow.

The idea for ainocode.top came during a lunch break. I'd just read an article about how non-technical people were launching businesses with no-code tools, and it hit me: I could build a website that curates exactly those kinds of guides—for people like me who want to use AI without writing code. The side goal was to one day earn a little ad revenue through Google AdSense. Maybe $50 a month to cover my coffee habit. That was the dream.

I didn't know if I could actually do it. I'd tried WordPress years ago and got stuck on hosting. But this time I had AI assistants that could literally write code for me. The moment I realized I could just describe what I wanted in plain English and watch a site appear—that was when I committed.

The Tools I Used (and Why)

Every tool I used was chosen for one of two reasons: it was either extremely cheap or completely free, and it had to work from China (sometimes with a VPN, sometimes without). Here's the stack:

**Namesilo for the domain**

I needed a domain that was cheap and could be paid for with Alipay. Namesilo ticked both boxes. ainocode.top cost me $1.88 for the first year. No hidden fees, no upsells. You don't even need a VPN because Namesilo loads fine from China and Alipay is built right in.

If you are ready to register your own domain, use my link:

Namesilo and enter coupon code **ainocode** at checkout to save $1.00.

**Cloudflare Pages for hosting**

Completely free. Unlimited bandwidth, automatic HTTPS, and a global CDN. It connects directly to a GitHub repository and redeploys the site every time I push a change. Setting up Cloudflare Pages took me maybe 20 minutes. There's no server to maintain, no bill at the end of the month. For a static site like mine, it's perfect.

**VS Code + Cline for AI-assisted building**

Visual Studio Code is the editor, and Cline is the extension that lets an AI agent write and edit code right inside the project. I would open Cline, describe the page or feature I wanted, and it would generate all the files. Then I'd review the changes and click "approve." I never wrote a single line of code by hand.

**DeepSeek API as the AI brain**

Why DeepSeek and not OpenAI? Two reasons: cost and accessibility. From China, connecting to OpenAI requires a reliable VPN, and the bills can add up fast if you're making dozens of API calls an hour. DeepSeek is based here, the API is dirt cheap, and I didn't need a VPN to use it. Over two weeks of building, I spent 14 RMB total—about $2. That's it. And the quality of the code it generated through Cline was shockingly good.

**Gemini for planning, Claude for reviewing**

Before building, I used Google Gemini to brainstorm the site structure, article categories, and SEO basics. It was like having a strategy consultant. Then, when Cline finished a chunk of work, I'd often paste the output into Claude (via the free tier) to get a second opinion on code quality. Claude caught a few CSS issues that would have broken the mobile layout. The combination gave me confidence that even though I'm not a developer, I wasn't just blindly accepting whatever the first AI said.

Step-by-Step: How I Built It

**Day 1: Buying the domain**

I went to Namesilo, searched for "ainocode," and saw that .top was available for just $1.88. I added it to the cart, selected Alipay, paid, and within five minutes I owned a piece of the internet. I didn't need a VPN, my Chinese ID, or anything special. It felt oddly empowering.

**Setting up Cloudflare Pages**

Next, I created a free Cloudflare account. I pointed my domain to Cloudflare's nameservers (Namesilo had a one-click setting for this). Then I went to the Pages section, connected my GitHub account, and selected the repository where I'd soon push my site's code. At this point the site was a "coming soon" blank page, but it was live at ainocode.top.

**Installing VS Code and Cline**

I already had VS Code on my laptop. I installed the Cline extension from the marketplace. Cline asked for an API key; I had already signed up for DeepSeek, topped up 20 RMB, and generated a key. I plugged it in, and Cline was ready. No command-line stuff, no npm installs.

**My building workflow**

This was the rhythm: I'd open Cline, type something like "Create a homepage with a hero section, a grid of blog post cards, and a newsletter signup form," and hit enter. Cline would think for a few seconds, then start creating files—index.html, style.css, maybe a JavaScript file. I'd see the code appear in my editor. I'd scan it (mostly looking for words I recognized), and if it made sense, I'd approve. Cline would then save the files, and I'd run the local preview in my browser.

When something looked wrong—like the navbar not sticking to the top—I'd tell Cline in plain English: "The navbar should stay fixed when I scroll." It would adjust the CSS and ask for approval again. This back-and-forth became a conversation.

**The first time the site appeared in my browser**

That moment is imprinted in my memory. I had a real, styled homepage with my own content placeholder text, running on my local machine. I almost called my wife over to look. For someone who'd spent years thinking websites were magic black boxes, seeing my own creation in a browser felt like I'd broken through a wall. I immediately pushed the code to GitHub, and Cloudflare Pages automatically deployed it. Within minutes, ainocode.top was live for the whole world.

The Real Costs, Broken Down

If you're still skeptical about the price, here's the exact breakdown:

| Item | Cost |

|------|------|

| Domain (Namesilo) | $1.88/year |

| Hosting (Cloudflare Pages) | $0 |

| DeepSeek API credits | ~$2 (14 RMB) |

| **Total first-year cost** | **~$3.88** |

Compare that to traditional web development. Hiring a freelancer to build a similar blog site would cost anywhere from $500 to $5,000. Using a no-code platform like Wix or Squarespace would run you at least $10–$15 per month, plus domain. And then you'd still have to design everything yourself. My ongoing monthly cost is essentially zero—just the domain renewal next year at a slightly higher price (around $5, still negligible).

💡 **Ready to start your own site?** I registered ainocode.top at Namesilo for just $1.88. Use coupon code **ainocode** to save $1.00 at checkout — the cheapest way to own a piece of the internet.

I deliberately tracked every expense because I wanted to prove to myself and others that money isn't the barrier anymore. The barrier is knowing that this path exists.

What Went Wrong (Honest Section)

It wasn't all smooth. I hit several walls, and each one taught me something.

**GitHub connection took multiple attempts**

When I tried to link Cloudflare Pages to my GitHub repository, I kept getting an authorization error. I'd granted the right permissions, but Cloudflare didn't see the repo. After an hour of retrying, I realized I'd created the repository under a different GitHub account. I had to delete and recreate the Cloudflare integration. A small but frustrating detail.

**Cloudflare deployment configuration confusion**

My site uses a static site generator, so Cloudflare needed to know the build command and output directory. I left those fields blank at first, thinking "I don't have a build command," but the deployment failed. I eventually figured out I needed to set the output directory to the root and the build command to nothing, but the error messages were cryptic. I had to search forums and ask Claude to translate the error into plain English.

**DeepSeek API key copied incorrectly**

I copied the API key from the DeepSeek dashboard, but there was an extra space at the end. Cline kept saying "Invalid API key." I retyped it three times, regenerated a new key, and finally realized I'd pasted it with a trailing space. A classic facepalm moment.

**AI sometimes generated code that looked right but broke on mobile**

Once, Cline added a sticky header that worked beautifully on desktop but covered half the screen on my phone. I didn't catch it until I tested on a real device. Now I always ask Cline to "make this responsive" and test early.

The lesson? Expect small obstacles. Every single one was fixable, and none required deep technical knowledge—just patience and the willingness to Google (or AI-ask) the error message. If you're the type who gives up after one failed attempt, this might be hard. But if you treat it like a puzzle, you'll get through.

What I'd Do Differently

Looking back, I'd change a few things to speed up the process and get better results.

**Start with a content strategy before building**

I jumped straight into coding the site structure without a clear plan for what articles I'd write. After launch, I struggled to produce consistent content. If I'd mapped out 20 article ideas and even drafted a few before building, the site would have been more valuable from day one.

**Write 10 articles before launching**

An empty blog looks sad. I launched with 3 articles, and it took weeks to fill the site up. I'd now advise anyone to write at least 10 posts in advance and schedule them. This also helps with SEO because Google sees a site with depth and freshness.

**Set up Google Search Console on day one**

I didn't submit my sitemap until two weeks after launch. That delayed indexing. Search Console is free and takes 10 minutes to connect via Cloudflare. Do it the moment your domain is live.

**Be more specific in AI prompts from the start**

Early on, I gave vague instructions like "make a nice footer." Cline would generate something generic. Later I learned to say, "Create a footer with three columns: quick links, social media icons, and a copyright line. Use a dark background with white text." The results were dramatically better.

Can You Do This Too?

Short answer: absolutely. If you can use WeChat, order a taxi on Didi, and follow a recipe, you have all the technical skills required. The tools handle the heavy lifting. Here's what you actually need:

• **An idea** – even a small one, like a blog about your pet or a portfolio for your photography.
• **Patience** – the first few hours will feel confusing, but the learning curve is shallow.
• **A willingness to talk to an AI** – you'll describe what you want, review what it builds, and correct it. That's the whole process.

I won't pretend it's zero effort. You will need to read a few guides, watch a short tutorial on setting up Cloudflare Pages, and spend an afternoon tinkering. But compared to learning HTML, CSS, JavaScript, and DevOps, this is like using a remote control instead of rewiring the TV.

The biggest barrier is believing you can. I'm a banker, not a programmer. My colleagues still don't believe I built ainocode.top myself. If I can, you can.

What's Next

The site is live, but the journey has just started. I'm now focused on three things:

First, growing the content library. I publish two articles per week, and I use AI (mostly Claude and DeepSeek) to help outline and research, but I always write the final draft myself. The voice has to be human.

Second, applying for Google AdSense. I need about 20-30 solid posts and steady traffic before I'll get approved. I'm aiming for that in the next three months.

Third, building more AI-powered side projects. This experience opened my eyes. I'm already sketching out a simple directory site for AI tools and maybe an automated newsletter. The cost to experiment is near zero, so why not?

If you've been sitting on a website idea, let this be your sign. The only difference between me and you is that I clicked "buy domain" and started talking to Cline. You can do the same tonight.

**Your first step:** grab your domain at Namesilo — I paid $1.88 for ainocode.top. Use coupon code **ainocode** at checkout to save $1.00. Then come back and follow the guides on this site to build the rest.