# Marcelo Retana

> Full-stack developer with 10+ years of experience building high-converting websites, web apps, mobile apps, and MVPs. Specializes in performance-first development with Next.js, SEO optimization, and AI search visibility. Based in Costa Rica, delivering worldwide.

## About

- [Portfolio](https://marceloretana.com): Main site with work experience, client testimonials, services, and contact information
- [Blog](https://marceloretana.com/blog): Technical articles on web development, AI developer tools, SEO, and lead generation

## Services

- MVP & Prototype Development: Launch-ready MVPs and demos for startups. Fast turnaround from idea to working product.
- Professional Websites: High-converting business websites with modern design, fast performance, and SEO optimization.
- Web & Mobile Apps: Full-stack web applications and React Native mobile apps built for scale.
- Ecommerce Development: Custom online stores, Shopify apps, and ecommerce solutions that drive sales.
- AI Workflow Integration: Custom AI integrations and automated workflows using modern LLM tools.

## Tech Stack

- Frontend: React, Next.js, TypeScript, Tailwind CSS, React Native
- Backend: Node.js, GraphQL, PostgreSQL, AWS, Docker
- AI & Tooling: Claude Code, MCP servers, AI-powered development workflows

## Blog Posts

- [Codex vs Claude Code: Dejá de Cambiar, Empezá a Dominar](https://marceloretana.com/blog/codex-vs-claude-code-deja-de-cambiar-empieza-a-dominar): Cada vez que sale una herramienta de IA nueva, los devs saltan de barco. El problema no es la herramienta — es nunca profundizar lo suficiente con ninguna. Acá va una comparación honesta y por qué dominar gana siempre.
- [Codex vs Claude Code: Stop Switching, Start Mastering](https://marceloretana.com/blog/codex-vs-claude-code-stop-switching-start-mastering): Every time a new AI coding tool drops, devs jump ship. The problem isn't the tool — it's never going deep enough with any of them. Here's an honest comparison and why mastery beats tool-hopping.
- [Qué Es llms.txt y Cómo Agregarlo a Tu Sitio Web](https://marceloretana.com/blog/que-es-llms-txt-y-como-agregarlo-a-tu-sitio-web): Los modelos de IA están leyendo tu sitio web ahora mismo — pero están adivinando qué es importante. llms.txt es un nuevo estándar que te permite decirles directamente. Acá te explico qué es, quién lo usa, y cómo implementarlo en Next.js.
- [What Is llms.txt and How to Add It to Your Website](https://marceloretana.com/blog/what-is-llms-txt-and-how-to-add-it-to-your-website): AI models are reading your website right now — but they're guessing what matters. llms.txt is a new standard that lets you tell them directly. Here's what it is, who's using it, and how to implement it in Next.js.
- [Por Qué Claude Code Se Come Toda Tu RAM (Y Cómo Solucionarlo)](https://marceloretana.com/blog/por-que-claude-code-se-come-toda-tu-ram): Ghostty mostrando 207GB en una Mac de 32GB. Memory leaks de 129GB en Linux. No son solo malos hábitos — hay culpables técnicos reales escondidos en tu sistema. Esto es lo que realmente está pasando.
- [Por Qué Tu Sitio Web No Genera Clientes (Y Qué Hago Diferente)](https://marceloretana.com/blog/por-que-tu-sitio-web-no-genera-clientes): La mayoría de sitios web se ven bien pero fallan en lo único que importa — convertir visitantes en clientes. Así es como el desarrollo full-stack moderno, la arquitectura enfocada en rendimiento y la optimización para búsqueda con IA generan resultados reales.
- [Why Claude Code Eats All Your RAM (And How to Fix It)](https://marceloretana.com/blog/why-claude-code-eats-all-your-ram): Ghostty showing 207GB on a 32GB Mac. 129GB memory leaks on Linux. It's not just bad habits — there are real technical culprits hiding in your system. Here's what's actually going on.
- [Why Your Website Isn't Generating Leads (And What I Do Differently)](https://marceloretana.com/blog/why-your-website-isnt-generating-leads): Most websites look fine but fail at the one job that matters — turning visitors into customers. Here's how modern full-stack development, performance-first architecture, and AI search optimization deliver real business results.
- [Cómo Multipliqué x10 Mi Velocidad de Código con Claude Code (Y Mi Setup Open Source)](https://marceloretana.com/blog/como-multiplique-mi-productividad-con-claude-code): Después de meses refinando mi flujo de trabajo con IA, comparto la configuración exacta que convirtió a Claude Code en mi herramienta más productiva.
- [How I 10x My Coding Speed with Claude Code (And My Open Source Setup)](https://marceloretana.com/blog/how-i-10x-my-coding-with-claude-code): After months of refining my AI-powered development workflow, I'm sharing the exact configuration that turned Claude Code into my most productive coding partner.
- [Por Qué Construyo MVPs Rápido (Y Vos También Deberías)](https://marceloretana.com/blog/por-que-construyo-mvps-rapido): La velocidad no es recortar caminos - es aprender más rápido. Así es mi enfoque para el desarrollo rápido de MVPs.
- [Why I Build MVPs Fast (And You Should Too)](https://marceloretana.com/blog/why-i-build-mvps-fast): Speed isn't about cutting corners - it's about learning faster. Here's my approach to rapid MVP development.

## Contact

- [Book a Call](https://cal.com/marcelo-retana): Schedule a free consultation
- [GitHub](https://github.com/maketroli): Open source work and projects
- [LinkedIn](https://linkedin.com/in/marceloretana): Professional profile
- Email: info@gexpsoftware.com