Your First Landing Page with AI: A Step-by-Step Walkthrough
Build and launch a polished one-page website using nothing but natural language prompts — no prior coding experience needed.
What we are building — and why one page is the perfect place to start
A landing page is a single web page designed to promote one thing: a product, a service, a hobby, or an event. Unlike a full website with multiple pages, navigation menus, and a database backend, a landing page lives in one HTML file. That simplicity makes it the ideal first project for anyone curious about building with AI.
A bakery advertises its weekly specials on one. A freelancer shows their portfolio on another. Each one was written and deployed by someone — and with modern AI tools, that someone could be you.
By the end of this guide you will have a live landing page at a public URL. You will have written zero code by hand. And you will have a repeatable process you can apply to any future project.
Step 1: Choose your AI assistant
You need an AI tool that can generate and edit HTML, CSS, and JavaScript. Here is how the most common options compare:
| Tool | Best for | Getting started |
|---|---|---|
| ChatGPT | Quick prototypes in the browser | chat.openai.com — no sign-up required to start |
| Claude | Longer, more detailed conversations | claude.ai — free tier available |
| Cursor | Editing code like a professional | Download the editor; AI features built in |
| Lovable | Full app generation from chat | lovable.dev — handles hosting automatically |
The prompts below work in any tool — the results will vary slightly but the process is the same.
If you have never used an AI coding tool before, Vibe Coding for Beginners covers the setup and mindset you need before your first session. For a broader foundation, A Simple Guide to AI Coding explains how to structure prompts for reliable results.
Step 2: Write your brief — the most important prompt you will send
The most important rule of AI-assisted building is this: describe the outcome, not the implementation. Tell the AI what you want the user to see and feel, not the technical details of how to achieve it.
Start with a short brief about your project. Here is a template you can paste into any chat:
Create a single-page landing page in one HTML file with embedded CSS.
My project is about: [one sentence describing your idea]
Design requirements:
- Modern, clean layout
- Mobile-responsive — looks good on phones and desktops
- A colour scheme that feels [energetic / calm / professional / playful]
- A hero section with a headline, subheading, and call-to-action button
- A features section with three columns showing key benefits
- A contact or sign-up section at the bottom
Use vanilla HTML and CSS only — no external frameworks or libraries. I want to open the file in a browser and see the full page.
Replace the placeholders with your own details. Here is a concrete example for a fictional home bakery:
Create a single-page landing page in one HTML file with embedded CSS.
My project is about: a home bakery called "Dave's Sourdough" that sells artisan bread and pastries in my local area.
Design requirements:
- Modern, clean layout
- Mobile-responsive
- A warm colour scheme — cream background, dark brown text, accents in terracotta
- A hero section with the bakery name, tagline "Real bread. Real simple.", and an order button
- A features section showing three columns: "Sourdough Starters", "Weekly Bakes", "Local Delivery"
- A contact section with an email sign-up form
Use vanilla HTML and CSS only — no external frameworks.
The AI will generate a complete HTML file. Save it as index.html and open it in your browser. It will not be perfect yet — that is what the next steps are for.
Step 3: Review your first draft like an art director
Look at your page and ask three questions:
- Does the layout match what I described? Are the sections in the right order?
- Does the design feel right? Colours, spacing, fonts — do they match the tone you wanted?
- Is anything missing? A section you expected but did not appear, or text that reads oddly.
Make notes as you go. You will feed these back to the AI in the next step.
Step 4: Refine with focused feedback
Now you become the art director. Paste your observations back into the chat. Be specific about what you want changed and why.
Good start. A few changes:
- The headline is too small. Make it at least 60px on desktop.
- The hero background should use a subtle gradient from cream to light terracotta.
- Add 2rem of padding around the whole page so text does not touch the edges.
- The "order now" button needs to stand out more — terracotta background, white text, rounded edges (border-radius: 999px).
- Replace the placeholder image in the hero with a simple bread illustration drawn with CSS or inline SVG.
Each change takes seconds. After three or four rounds of feedback, the page will start to feel polished.
Step 5: Add the essential landing page sections
A great landing page follows a proven structure. Here is the anatomy you should aim for:
| Section | Purpose | Key elements |
|---|---|---|
| Hero | Grab attention immediately | Headline, subheading, primary button |
| Features | Explain what you offer | Three-column grid, icon per item |
| Social proof | Build trust | Testimonials, client logos, user count |
| Call to action | Drive the next step | Sign-up form, order button, contact link |
| Footer | Wrap up cleanly | Copyright, social links, privacy note |
Add each section one at a time through separate prompts. Building incrementally — rather than asking for everything at once — gives the AI fewer chances to misinterpret your intent.
For example, to add a testimonial section:
Add a testimonials section between the features and the call-to-action area.
Include:
- A heading that says "What our customers say"
- Three testimonial cards in a row (stack on mobile)
- Each card has a quote, the customer's name, and a star rating
- The cards should have a white background with a subtle shadow and rounded corners
Use realistic-sounding placeholder text for the quotes.
Step 6: Deploy in two minutes
Your landing page is a single HTML file. Deploying it is free and takes two minutes.
The simplest option is Netlify’s drag-and-drop deploy. Go to the site, drag your index.html file onto the browser window, and Netlify gives you a public URL instantly. No command line, no configuration, no credit card required.
If you used Lovable, it handles hosting automatically with every save. Cursor and Claude users can deploy through Vercel or Netlify the same way.
Once you have your URL, share it. Send it to friends, post it on social media, or add it to your email signature. A live URL is the finish line — and you crossed it without writing a line of code.
You now have a repeatable process
Brief, prompt, review, refine, deploy. That loop works for a landing page today and for a multi-page site or web app tomorrow. Every time you go through it, you will get better at spotting what to include in your prompts and what to fix in review. The AI handles the syntax; you handle the decisions.
For a comprehensive guide to writing better prompts and building more complex projects, see A Simple Guide to AI Coding. If you want to understand how the generated code works under the hood, Vibe Coding for Beginners breaks down the patterns that produce reliable, production-quality output.
Further reading
Learn to use AI tools like Claude Code, OpenAI Codex, Cursor and Gemini to build websites, apps and software using plain language prompts.
View book
How to use AI and LLMs to create apps, websites and amazing technology — without needing to be a programmer or write code from scratch
View bookMore insights
All ArticlesA 90-Day Roadmap for Rolling Out Agentic Coding on Your Team
Roll out agentic coding without the chaos — a phased plan that gives your team standards, governance, and a rollout pace they can actually sustain.
Read articleContext Windows, Memory, and Why Your AI Keeps "Forgetting" the Project
Discover why your AI loses track of your project mid-session and how context windows, memory files, and smart prompting keep conversations on track.
Read articleCursor Rules, .cursorrules, and Project Context That Sticks
How a single .cursorrules file keeps your AI coding tool grounded in your project's stack, conventions, and preferences — session after session.
Read articleCursor vs Claude Code vs GitHub Copilot: A Practical Comparison for 2026
See how Cursor, Claude Code, and GitHub Copilot compare across real workflows — and which AI coding tool fits your team and project in 2026.
Read article