Why I built Unslop

I've always liked writing. Blog posts mostly, the kind where you actually have something to say.

Like a lot of people, I started leaning on AI to help. English isn't my first language, so I used it to catch typos and small mistakes, and to reach for words I wouldn't have found on my own.

However I used it, the text came back sounding like AI. Once you've read enough of it, you start to see the shape: the same rhythm, the same turns of phrase, again and again. It got under my skin.

I wanted something that would still tighten my writing, even lend it a wider vocabulary, without making it read like every other AI page on the internet.

The scale of it makes the problem obvious. Now that anyone can produce text by writing to agent, you open LinkedIn or a company blog and it all sounds the same. The same filler, the same tics, the em-dashes everyone has learned to spot. Work with these models long enough and you stop being able to unsee the patterns.

There's an unfair edge to it too. Search engines now push down anything they think a machine wrote. For low-effort write-me-a-blog-post content, fair enough. But plenty of people put real work into an article and use AI only as a final polish, and their writing gets flagged the same way. Unslop is for them.

A bit about me. I'm a software engineer. I've built startups for fifteen years, and I still like shipping products more than anything. This is one of them.