Saying that this has been quite the ride would be a gross understatement. Siteimp started life as a bunch of scripts that I used in my ops practice. Originally those scripts were purely about web application performance and in the earliest phase they would do stuff like log in, do a thing and collect metrics. I learned a lot about building fast applications and then, my own websites weren't quite as fast as I wanted them so I modified the scripts, build a really fast scraper and started doing full website performance tests.

I had plans for Siteimp but they just didn't quite happen. Whenever I would build a cloud based system it would end up looking like a ddos tool and keeping that beast tamed took up more time than adding features. Through development, I kept running into problems that felt insurmountable and likely were. Even things like billing were complicated and evolved into a complex meter based system that reminded me more of fantasy baseball than accounting.

And so Siteimp was put on pause. But then, stuff started to change. For one, I started building desktop software. Electron didn't quite solve my original problem, but a rewrite revealed that Tauri did. And then, I was working on the Fitness Tracker website when I started running into problems.

  • I had no good way to look at my overall content architecture or how pages were linked together and outside of the website.
  • I had no easy way to find all my broken links.
  • Running full website accessibility tests was a real pain.
  • Even things that should be relatively simple like analyzing meta descriptions alongside my title and heading structure were a real pain.
  • Finding dead pages was a weirdo blend of Puppeteer scraping and xml parsing.

It annoyed me...but Siteimp (a piece of software I had not committed any code to in four years) solved most of those problems. It was four years old, didn't quite match my current views on content, accessibility and performance but it was still pretty close.

That was on February 21, 2026. By February 24, I had planned out the project, broken it down into three hard sprints and made my first commits. Siteimp-desktop was officially born.

It took almost five sprints, but when I build alone, I can get obsessed with quality. In all honesty, I really didn't expect to write a full in app support system based on Formimp. My previous three bot system evolved into a five bot system and I added on a media bot and external link checker bot. The old synthetic based system had to go, but it was replaced by something even cooler that I'll be releasing in the next couple of months. When I started, I thought I'd do a really basic EXIF implementation so that people who publish their own photos to the web could be sure they weren't leaking information they didn't want to, but that turned into a full ExifTool implementation.

But the biggest drain on time was that at each step of the process, I kept finding horrifying problems on my own websites. From open graph images that didn't exist to another two that actually thought they were text because I can make unbelievably stupid decisions. I had my own Office Space moment in which I realized that I was good enough to build software that could prove beyond a doubt that I often fuck up the small things.

I got high on my own supply.

But, Siteimp is ready for beta. It currently builds for Windows only, but there are vague plans to fix that. Those plans are vague because I'm a better designer now than I was a year ago when I decided that I needed to learn. But I'm stuck in this weird stage where I'm good enough to see that Apple applications have a unique visual language and expected functions while not being quite good enough to figure out the concepts behind the functionality. Rather than release utter shit, I can be patient and make sure that Apple users get the experience they deserve and expect.

And here's the bonus of all of that. Now that Siteimp is ready for beta, I can start fixing all those ugly little monstrosities that became test cases. The saga continues.