Website Relaunch with AI Instead of an Agency: What Used to Cost €15,000 Now Takes 2 Days

In early February 2026, I faced a decision: my website needed a relaunch. New design, better structure, more professional presentation. The classic route would have been: brief an agency, collect quotes, wait 4–8 weeks, invest €10,000–20,000.

I chose a different path. And this article describes honestly how it went.

The Tech Stack: Hugo, Claude, and OpenCode

Hugo – the Static Site Generator

Hugo is an open-source framework that generates websites from Markdown files. No database, no WordPress, no PHP. Just HTML, CSS, and JavaScript – lightning fast, secure, and easy to host.

This isn’t a new tool. Hugo has been around for years and is used by thousands of companies. What matters is: The combination with AI turns Hugo into a turbocharger. Because Markdown is text – and text is exactly what AI models process brilliantly.

Claude – Anthropic’s AI Model

Claude is a large language model that can write, analyze, and refactor code. Not theoretically, but practically: CSS files with 2,000+ lines, Hugo templates with Go templating, responsive layouts, SEO optimization.

OpenCode – the Bridge Between AI and Codebase

OpenCode is a CLI tool that connects Claude directly to my project directory. It can read, edit, and create files, run Git commands, and even trigger deployments. This is the crucial difference from ChatGPT in the browser: OpenCode doesn’t work with copy-paste snippets – it works directly in the project.

What We Actually Built

Design and Structure

• Complete redesign of all content pages: Services, Workshops, References, Whitepapers, About, Contact
• Card grids instead of walls of text – visual structure that works on desktop and mobile
• Accordion elements for details that don’t need to be immediately visible
• Stat cards and result cards on the About page
• Contact grid with booking, email, and phone as standalone cards

Technology and Infrastructure

• Matomo analytics (self-hosted, cookieless, GDPR-compliant) – configuration-driven from hugo.yml, active only in production
• Bilingual support (German/English) via Hugo’s filename-based i18n
• SFTP deployment to Strato hosting via Python script
• Git-based workflow with GitHub as repository

Content and SEO

• All copy revised and restructured
• Meta descriptions and Open Graph tags for every page
• Semantic HTML with correct heading hierarchies
• Performance: No external dependencies, no JavaScript frameworks, no jQuery

What the Workflow Looks Like

Here’s how a typical work session runs:

1. I describe what I want – in plain language. “The contact page should have three cards: booking, email, phone. Below that, accordions for details.”
2. OpenCode analyzes the existing code – reads the relevant files, understands the structure, checks CSS variables and layout patterns.
3. Claude writes the code – Markdown content, HTML structure, CSS styles. Everything matching the existing design system.
4. OpenCode applies the changes – directly in the files, no copy-paste.
5. Build, commit, deploy – Hugo builds the site, Git saves the state, SFTP pushes everything to the server.

The entire relaunch – design, content, technology, deployment – took about 2 days. Not weeks. Not months. Two days, spread across multiple sessions. Individual changes go live in minutes.

The Honest Math

What an Agency Would Have Charged

What It Actually Cost

In fairness: my time has value too. And the API costs for Claude are real – the model charges per token, and a project of this size adds up to several million tokens. But even with everything factored in, I’m at a fraction of agency costs – and I have full control over every aspect of the website.

What AI Can Do – and What It Can’t

What Works Surprisingly Well

• Writing and debugging CSS – Claude understands specificity conflicts, CSS variables, responsive breakpoints
• Hugo templates – Go templating, conditionals, i18n logic
• Maintaining consistency – new components match the existing design system
• Finding bugs – “The button text isn’t readable” → Claude finds the CSS rule causing it and fixes it
• Deployment automation – SFTP scripts, Git workflows

Where Human Judgment Remains Essential

• Strategic decisions: Which pages do I need? What’s the core message?
• Tone of voice: Does the text sound like me or like generic AI output?
• Visual feedback: Does this actually look good on screen?
• Business priorities: What needs to go live first?

AI doesn’t replace thinking. It replaces typing, Googling, and hours of debugging.

What This Means for SMEs

If a solo consultant can build a professional, bilingual website with analytics, SEO, and automated deployment in a few days – what does that mean for companies with 10, 50, or 500 employees?

The implications go far beyond websites:

• Internal tools can be prototyped faster
• Documentation can be created and maintained more efficiently
• Automation scripts are built in minutes instead of days
• Digital projects need less external budget

This doesn’t mean agencies become obsolete. For complex web applications, e-commerce platforms, or enterprise systems, professional development remains indispensable. But for corporate websites, landing pages, and content platforms, the cost-benefit equation shifts dramatically.

Bottom Line

This website relaunch isn’t a tech experiment. It’s proof that AI-assisted development is production-ready today – if you know how to use the tools.

The website you’re reading right now was built this way. Every page, every CSS element, every deployment. Not in weeks with a five-figure budget, but in hours with an AI copilot.

This isn’t the future. This is February 2026.

Want to know how AI can accelerate your projects? Let’s talk about it – pragmatically, without hype.

→ Schedule a call

Tools: Hugo · Claude by Anthropic · OpenCode