Tools sometimes have limits in terms of what they’re capable of.

Case in point: Have you ever tried to remove a hydraulic lift from an office chair? It is the most painful process, one that seems designed to wear down anyone that tries to do it. I do not recommend it.

If you look around for more than five minutes, you will see some expert on this topic explain how they do it easily. They tell you to do things like use WD-40 or use a pipe wrench, but you try those things and nothing seems to happen. (For some reason, they often use example chairs that look good on camera, rather than the ones that actual people are struggling to fix.) Loved ones ask you why you’re making so much noise and you’re embarrassed to explain it’s because you’re tying to remove a broken hydraulic lift from an office chair.

Today, I saw some company that had clearly gotten so sick of this problem that they developed a “Combination Universal Gas Cylinder And Seat Mechanism Kit”—essentially a hydraulic lift that’s already connected to the part that’s impossible to remove. And they charge $90 dollars for it, at which point you might as well start looking for a new chair.

(Free idea, inventors: A version of an office chair hydraulic lift that can be removed by mere mortals without having to own a metal shop. I should see if that guy who developed all those Kensington locks might be interested.)

My personal problems with office chairs aside, I do think it highlights something that is very much true about the way that we build things—we often get caught in routines that just feel like they’re not getting us anywhere, and they just leave our brains feeling like mental jelly.

But recently, I discovered something that made me feel creative again after a long break: I loaded up a couple of tools that took different approaches to solving problems than I was used to, and all of a sudden, I felt like some ideas unlocked themselves that I’d had trouble approaching previously, in part because they presented things in ways that weren’t quite in the same wheelhouse as I had experienced previously.

Penpot is secretly one of the best FOSS webapps we have and we should build more things like it.

One of those things was Penpot, an open-source design and prototyping tool that you can self-host on a computer. A good comparison point to it is Figma, but the ability to build things in a web browser, on a server you own, and for it to look and feel incredibly polished, I think shows me that there’s a real potential for open-source self-hosted software as a whole to get a lot better.

As you might remember, earlier this year I wrote a long feature about self-hosted tools, where I considered what was good and was was really questionable. (On the other hand, I still have problems with NextCloud.) It has been interesting to see some of these tools continue to evolve. Activepieces, the FOSS take on Zapier I gushed about, has significantly improved to the point where it might arguably be better than Zapier ever was, and it did that all in the span of like, a year. There was clearly a demand for a Zapier-like tool that didn’t cost tons of money to operate.

But honestly, the tool that I’m most impressed by that I’ve tried lately is Directus, which is essentially a no-code backend tool. It’s often put into the headless CMS category because that’s where it tends to most obviously fit, but it can literally be used to collect, manage, and analyze any kind of data—something highlighted by the fact that the most recent product by its parent company, Monospace, is essentially a framework for the platform, something called AgencyOS.

For people with a lot of stuff to manage, it feels like glue—and even better, it feels comparable to technology tools that software-as-a-service companies charge lots of money for. And while its license is arguably not pure FOSS—instead, it uses a generous Business Source License to ensure that only companies with less than $5 million in finances leverage it in an open-source way—it is closer to the open-source ethos than many other companies playing in this space are. (And you’ll notice that after three years the software reverts to GPL—which is something that did not have to do.)

Directus is doing complex things, but its design doesn’t feel complex. It feels approachable, and is a great example for other companies building self-hosted things to emulate. And it has amazingly detailed video tutorials for everything, making it perhaps the friendliest onboarding experience I’ve ever seen for a self-hosted tool.

So as a known CMS nerd, I didn’t give it a chance until now. Why? Well, at the time that I last moved content management systems, it was a PHP-based data management system—but a couple of years ago, they made a big bet to rewrite the whole thing in up-to-date technologies such as TypeScript and Node.js, which resulted in a polished, improved, result that not only looks great, but runs very effectively as well. It passed the Docker test in spades, too—I got the whole thing to run cleanly with a single lightly adorned config file.

Finding new tools and having them work the way you expect is freeing, because it means that you’ve unlocked something in yourself that you didn’t think you had before. Being able to use Directus and Penpot has started to open up some fresh ideas in ways I was missing previously, and I really appreciate that. They make me see possibilities.

Don’t just keep beating on a broken office chair, hoping for something to change. Find a new way to think, and eventually, the creativity will start flowing.