I Tried Hostinger’s New Horizons AI Developer Tool: Is the Hype Justified?

The tool was initially marketed as an easy way for non-developers without a technical background to build their own apps, but these days Hostinger has been leaning more into its use as an AI website builder. I used it to build both an app and a website, and I will share the results with you. I’ll also provide my overall thoughts on Hostinger Horizons’s strengths, limitations, and usefulness. But first, let me give you a brief overview of the service, its main features, and how it works.

Hostinger Horizons: at a glance

How to use Hostinger Horizons

To begin, go to the Horizons website:

As you can see in the screenshot above, getting started with Horizons is as straightforward as using ChatGPT or Claude. I would argue that it’s even more straightforward because there aren’t any models to choose from or options to tweak. It’s nothing more than a chat window where you describe what you’d like to build.

If you have samples or example images you’d like Horizons to use for inspiration, you can attach those with your prompt. Either way, it gets to work immediately.

This “take it and run with it” approach is actually something that I didn’t like when I started using it. I even specifically instructed it to not start building right away, but it ignored me and did so anyway.

Despite being initially mildly annoyed, it turned out to be only a minor issue. Nonetheless, I would have preferred to map out my idea before getting started.

I think the tool could benefit from having an optional startup wizard. Especially since it’s primarily aimed at non-developers, who might have great ideas, but probably also a lot of blind spots. A wizard could help them flesh out their idea in greater detail before moving into the “build it” phase.

An overview of my experiments with Hostinger Horizons

Altogether, I ended up running two experiments where I built the following:

• A somewhat complex time-keeping app with advanced features.
• One page website with interactive scroll effects.

1. Time-keeping app with advanced features

I began with the more complicated of the two challenges because I was curious how far I could push the tool before I started hitting speed bumps. It did eventually happen, but overall, I managed to get quite far in only a few hours. Here are the features that I added to the app:

• Dashboard with various pieces of information that are pulled from other areas of the app
• Time tracking that lets you start a timer, pause it, and end it to complete a shift. You can also manually add shifts and edit existing shifts. In addition, there is a toggle that switches between a list view of your shifts and a calendar view.
• Clients and projects area where you can add clients, set your pay rate per hour, as well as the currency you’ll be invoicing in. You can also create projects under each client, which can help you track what you’re working on.

• Invoices that are customizable with your logo and that you can export as PDF files.
• Profile section where you add your personal and business information, which then populates into the invoices.

What went right 😁

Right off the bat, from the very first iteration that Horizons created, the aesthetics were on point.

Although I didn’t include anything about colors, fonts, or how I wanted the app to look in my initial prompt, the result I got back was very modern and attractive. It seemed to lean in the style of what the Hostinger website might look like in a dark theme:

Beyond that, I was able to fix some initial bugs with prompts and I got the app to a point where it seemed to work. The buttons did what they were supposed to – at least it seemed like it at first glance. I could fill out all the fields, make my choices, export my made-up invoices as PDFs, etc.

However, upon doing some more comprehensive testing, I realized certain features didn’t work properly. The first glances were misleading.

What went wrong ☹️

The biggest bug that I discovered was that the core feature of my app didn’t work. That is to say, it didn’t track time properly.

I was able to start the timer and watch it tick away, second by second. I was also able to use the pause and stop buttons. However, if I went to another tab and left it on – which you’d do in a real-life scenario – the actual time that passed when I went back to it was way more than what the timer showed.

For instance, I left it on for about an hour, but when I checked on it, it showed that only nine minutes had passed.

There were also some other bugs with certain data not being pulled into the dashboard.

In defense of Horizons, I did not attempt to fix these bugs with further prompts because I wanted to move on with the other experiment.

I also felt like for the amount of time I put into it (about three hours in total) and what I got in return was incredibly impressive. Perhaps not deployable just yet, but definitely mind-blowing. Especially when you consider that five years ago, it would have taken a professional developer a few weeks (if not longer) to get the same prototype.

And if I was a developer working on this as a serious project, I could export the code and manually fix the bugs myself.

2. One page website with interactive scroll effects

For the website experiment, I decided to rebuild my own portfolio site. I wanted to add some interactive scroll features to it that I saw on another one page website.

I initially attempted to paste the entire source code of the inspiration site into Horizons. However, I ended up going over the character limit by ~63,000 characters! 🙈 To work around this limitation, I took screenshots of the source code instead. I then fed those to Horizons in a few batches:

With each batch, it extended the sections and scroll effects of the site.

When I had enough sections built, I gave Horizons a screenshot of my existing, live site. I asked it to extract the text and use it to rebuild the site in the style of the template we had been working on up to that point.

Afterwards, I prompted it to substitute all the placeholder links and images with the real ones from my site. I also asked it to correct some portions of text that it transcribed incorrectly.

When all was said and done, I was able to successfully recreate the one page website from scratch. It took me only 16 prompts!

What went right 😁

The biggest thing that went well was that Horizons was able to rebuild the site and incorporate the requested scroll effects from the inspiration website. They weren’t identical, but similar, and appealing in their own way.

I liked the result so much that I even exported the code because I will probably use it in some capacity:

What went wrong ☹️

Although Horizons did hit some minor bumps in the road on the path to the finish line, I was able to fix all but one of them.

It was such a minor detail that I only bothered to try to fix it once. When it didn’t work, I asked Horizons to simply delete it. The “it” was two words that had strike-through text. For whatever reason, Horizons couldn’t get the strike-through effect to render.

Other than that, everything else that went wrong was temporary and fixed with follow-up prompts.

So is Horizons worth it?

In my opinion, Horizons is a game-changer in many ways, and has something to offer people of all technical backgrounds.

• Non-developers with absolutely no technical background can use it to build simple applications and custom websites with relative ease. You’ll need to make some adjustments along the way, but the destination will absolutely be reachable.
• Non-developers with some technical chops will be able to get even further, building more complex apps and sites. If you understand how to use the inspect tool on your browser, this alone will take you far when hitting a wall with Horizons.
• Developers may even benefit the most from using Horizons – despite the tool not being marketed at them. Going from idea to prototype used to take weeks or even months, but with Horizons, the time is cut to mere hours. The fact that you can export the code and then modify it yourself is what makes this such a valuable tool. It would be even better if you could import code into Horizons as well, but perhaps Hostinger will add that in the future.

Overall, I believe that Horizons is a valuable service with unique strengths that go beyond what either ChatGPT or Claude can give you in response to coding prompts. As long as you are realistic about your expectations, maximize your prompts, and have some patience, you can use it to build some really cool projects.

Horizons pricing

Horizons has four plans, though they look different, depending on whether or not there is a promotion going on. More on that in a moment.

The standard entry-level Explorer plan starts at $6.99 per month and gives you 30 AI credits that renew every month. 30 credits might not seem like enough, but remember that it only took me 16 prompts to rebuild my portfolio site and 18 prompts to build an initial version of my time-keeping app. Granted, the app had bugs, but the point is that you can absolutely get something out of those 30 AI credits by taking a smart approach.

The only thing to be mindful of is that starting in the second year, you’ll need to pay for a Hostinger hosting plan to go with your Horizons plan. The combination can be more than double the cost of your first year contract.

In addition to the standard Horizons pricing you see above, Hostinger runs promotions from time to time. Depending on when you decide to head to the Horizons page, you may see something like this instead:

This promotion lets you try Horizons with 5 AI credits, which is enough to give you a feel for how it works. If you end up buying a plan, it also gives you a free head start on building your project.

Have you tried Hostinger Horizons yet? What did you think of it? If you built something interesting, share it in the comments below. I’d love to check it out.