I built an app with AI. It felt magical.
My honest experience with Lovable + how you can build your own app without coding
Hey friends,
A couple of days ago, I published my video collaboration with Lovable which I had a ton of fun working on. So, I’m dedicating this newsletter to share what I built and how I built it so you can try it out yourself.
Enjoy,
– Fawzi
How AI has changed my design process
Before Lovable reached out to me, I had already started experimenting and tinkering with their tool to build UX prototypes and other website/app ideas in my head. I’m using it to visualize and get feedback for the new educational site I’m building (you can learn more and join the waitlist).
Lovable has completely changed the speed at which I can bring my ideas to life and get early feedback them. As a designer who spends way too much time in Figma, it’s easy for me to get caught in the details and perfecting each pixel too soon. The details matter, but definitely not when you’re just starting off with a product that you’re not sure people even want. I’m still using Figma to create polished products and designs, but I finally feel like I have a “bicycle for the mind”.
For our collaboration, Lovable challenged me to build an app people could use. It took me a few weeks to come up with a useful idea, but lightning finally struck last weekend. It only took a few hours, and I had my web app up and running.
I’m calling it “never bored”, a simple site where you can browse events happening in your city today based on one criteria: your mood. It’s not a revolutionary app, but it answers the simple question (or job-to-be-done 😉): “What’s something fun I can do in my city today?”.
or just type neverbored.ai in your browser to use it!
We all need more fun in our lives
I got the idea for “never bored” during the weekend while I was browsing local events in Toronto. There are plenty of event aggregators to look at, but the lists were too long and some websites are littered with ads.
Too much noise.
I just needed a few options to pick from, instead of being overwhelmed by choices. For my app, I made a few assumptions about how people find fun events to attend:
People look for something to do based on their current mood (musical, artsy, active, relaxing, entertaining, etc.)
People don’t want a list of 200 events to scroll through
People want to find something without navigating multiple websites and pages
People want to spend more time outdoors than being indoors on their phones or TV
I could be wrong about all of these, but they’re good enough assumptions to start from.
Building “never bored”
The interface
I quickly designed an imperfect version of the homepage in Figma, then exported the screen as an image into Lovable. Then I prompted Lovable to build a responsive web app based on my design, along with general instructions of how I wanted it to work.
Building the interface was surprisingly easy. It stayed true to the original design I provided, and I could easily make tweaks to colours and fonts through prompts.
In the spirit of keeping it simple, I just needed three pages:
Homepage: collect the user’s input for event type/mood (musical, artsy, entertaining, educational, active, relaxing) and city
Loading: a short loading screen while the tool searches and summarizes events based on your criteria, with a dash of humour
Results: a results page showing 5 events in your city today based on your criteria, and each one would have a brief description, price, and a link to learn more
Showing only 5 results was an arbitrary decision, but it’ll help people choose something faster by keeping them focused on a few options (Hick’s Law). The goal is to get people to the “fun” faster even if results are imperfect. I hope it pushes people to get out of their comfort zone and attend something unusual.
The functionalities
While I had a smooth experience with the interface, I was worried about how to integrate the event search functionality in the background and I had low expectations of what AI tools could do with any sort of back-end development.
My site needed a search mechanism that would find events online and summarize them based on the user’s inputs. I had no idea how to do this so I just asked Lovable to see how it would respond.
Not only did it recommend using Perplexity’s search API, it provided the links for me to create an API key on my Perplexity account and another button to securely add my API key to my site (screenshot below). I loved this interaction because it gave me shortcuts to complete the actions I needed without having to switch to go in a research rabbit hole on Google or any other documentation.
Shortly after, I ran into a technical bug that killed my momentum because I had no clue how to resolve it. The connection to Perplexity’s API was failing and I was getting cryptic error messages that I couldn’t debug on my own. I tried googling the error message and couldn’t find any solutions.
As my hopes of building “never bored” were dwindling, I finally decided to just paste the error message into Lovable.
9 seconds later, the issue was resolved.
Lovable browsed the internet to look at Perplexity’s docs, discovered that the initial implementation it created used an outdated model, and updated it to the correct one.
“never bored” was finally working.
One last thing I wanted to try out was tracking usage analytics beyond the simple metrics like number of visitors, time on page, and bounce rate.
I wanted to know what people were searching for, what “moods” were the most popular, and which cities were most in demand. I prompted Lovable to create a table in my database to store what users were searching for and the results they were getting, without storing any personal or sensitive data tied to them.
Once again, it worked. In just a couple of days, “never bored” has registered 119 searches from across all continents!
AI is a tool, not a value prop
Besides having the neverbored.ai domain, I don’t mention AI once on the website. The truth: people couldn’t care less if my product (or yours) is “powered by AI” or not. They just want something that works and solves a problem in their life.
Someone recently said:
“Saying your app is powered by AI is like saying your TV is powered by electricity”
What you build matters more than the tools you use to build it, like I talked about extensively in the “AI-first” trap.
Try Lovable for free
If you haven’t tried Lovable yet, I highly encourage you to spend an hour building something. They have a free plan, and they’ve offered me a 20% discount to share with you for your first month on the Pro plan. You can redeem this discount by using FAWZI20NL on the checkout page.
If you build something, let me know and send me a link for me to try it out.
If you try never bored, send me your feedback because I’m going to keep building and improving it for fun.
Disclaimer: I don’t have any control over what specific events appear in the results, and it’s based on the results that Perplexity returns. Based on my testing, searching for the same mood and city twice will give slightly different results, which is good.
Share this post with someone
Share this post with friends, family, and coworkers.
If you’re not a free subscriber yet, join to get my latest work directly in your inbox.
⏮️ What you may have missed
If you’re new here, here’s what else I published recently:
You can also check out the Year 2049 archive to see everything I’ve published.
So cool fawzi. Tried your app - crazy you built that with AI 👀