The other day I just wanted to make a recipe for pasta carbonara.
What I got: three pop-ups before I could read anything, a life story about a trip to Italy in 2007, a blogger explaining how this recipe changed her life, and dozens of ads scattered through the text. The "Jump to recipe" button didn't work. So I scrolled, with greasy cooking hands, across my phone. And then, somewhere near the bottom: the recipe.
This is broken, I thought. And because building software is my job, the next thought came right behind it: I can fix this myself.
It started as something just for me
The first idea was small. One line: paste a link from a recipe site, get a clean recipe back. No ads, no stories, no pop-ups. Just the ingredients and the steps.
It wasn't meant to be a product. It was a side project for myself, for my own kitchen. My very first notes say it plainly: phase 1 was personal use. Whether anyone else would ever cook with it was not something I worried about yet.
But something else was going on too. I build software for a living, and the craft is changing fast right now because of AI. I wanted to know: how far do I actually get these days? If I have an idea I fully understand, how fast can I build it with AI as a helper? Parsely became as much an experiment as a tool.
It grew under my hands
Because I was cooking with it myself, I naturally noticed what was missing.
A recipe for four people, but I'm cooking for two? I wanted one click, no math. An American recipe in cups and Fahrenheit? I wanted grams and Celsius. And I wanted to save my recipes in lists: weeknights, parties, meal prep.
So every annoyance became a feature. What started as a kind of ad blocker for recipes grew into something that genuinely helped me in the kitchen. Not because I was following a grand plan, but because I was my own user. I built exactly what I was missing.
Built with AI
Then the experiment. Because building is my craft, the question wasn't whether it could be done, but how fast.
The short answer: surprisingly fast. The first working version of Parsely was up in two days. Not a prototype running only on my laptop, but a real first version that could fetch a recipe online and show it clean.
That was because AI took on the work I usually spend the most time on. The standard pieces, setting up the project, the boring but necessary work with no creativity in it. That left me time for the things that actually matter.
Because not everything happened on its own. A few things stayed handwork and my own decisions:
- The architecture. How everything fits together, which parts live where, and how it scales as it grows. That's thinking you have to do yourself.
- Fetching recipes. Pulling recipes apart cleanly from hundreds of different sites is harder than it sounds. That was the real puzzle.
- Taste and finish. How something feels, whether it's calm and clear, whether it's right. That stays a human choice.
The honest conclusion: AI made me, as an experienced developer, much faster, but it didn't replace the craft. It took the brakes off. An idea that used to sit in a drawer for months was now up and running in two days. That, I think, is the real shift in building right now.
From one website to everywhere
At some point I showed it to someone. And they wanted to use it themselves. Then someone else did.
That was the moment it became more than a side project. My own notes already had a phase 2: a public platform. So I took that step. What started on one website was extended to a mobile app, so you could save a recipe straight from your phone. And to a browser extension, so you could save a recipe right on the site where you found it.
Three ways to do the same thing: turn a messy recipe on the web into a clean recipe in your own library. It got more serious than I had planned. And I was fine with that.
What I learned from it
Two things stay with me.
The first: the best products solve a problem you have yourself. I didn't need to research what home cooks wanted. I was the home cook with greasy hands, scrolling.
The second: AI has changed how fast you go from an idea to something that works. Not by taking over the craft, but by driving the boring miles for you. What you do with that is still up to you.
Parsely started as an evening of frustration and a curious developer. By now it's a real place where you keep recipes from anywhere, clean and ad-free.
Want to build something great?
Parsely shows what's possible when you combine an experienced developer with the speed of AI. A first working version in two days, then built out step by step into a real platform.
Do you have an idea you keep coming back to, or a product that needs to be built better and faster? That's exactly what I do. I work as an agentic developer: I put AI to work to reach something functional fast, and I bring the craft to turn it into a thoughtful, maintainable product — from first idea to live platform.
Want to spar or use my help and insights for your own project? Feel free to get in touch. I'm happy to think along with you.
Try Parsely yourself
Want to see what came out of it? Try Parsely for free and save your first recipe — no pop-ups, no life story, no hassle.