WhenToTravel.com

I built two Claude Code skills that take a story from Azure DevOps ticket to merged PR. The operational overhead of solo development just got a lot smaller.

I had hundreds of content files across multiple languages and no way to read them all. So I built three AI-powered quality checks and a feedback loop that cut my review pile from hundreds of files to a handful.

I needed 350 content files for 70 destinations across 5 languages. One person, no budget for writers. Here's how I used a two-pass AI approach to generate structured travel content that didn't read like AI slop.

The use of Next-Intl's experimental useExtracted() feature combined with the Anthropic AI API allows for the automation of a website's internationalization workflow across multiple languages. This approach streamlines translation key generation, utilizes AI for context-aware translations, and integrates into development workflows via pre-push hooks.

When I first decided to internationalise whentotravel.com, I assumed the hard part would be translation. It turned out the real challenge was choosing the right approach for a content‑heavy Next.js site without over‑engineering myself into a corner. This post is a walkthrough of the dead ends, false starts, and eventual solution that finally clicked. This Ai image is so funny I had to include it. What kind of monitor / laptop combination is that! Omg! The Problem with the Of

I’ve set myself a few clear aims for 2026: Make something useful Learn something new Make some money doing it — ideally passive income On the learning side, I want to focus on: Productivity enhancement using AI Stronger, hands-on experience with modern UI technologies Getting much more familiar with embedding AI into a real product or tool I’ve owned whentotravel.com for quite a long time. When I bought it, the motivation was fairly simple: travel content has highly profitab