[Blog Pipeline #2] Killing the Machine to Build a Diary

Sometimes the most advanced system is the one you rip out.

From Automated Engine to Personal Log

I just tore this blog, Kevin’s Archive, down to its foundation. The old version was a machine, plain and simple. It had one category reserved for my own writing, while the rest was a network of automated pipelines designed for one thing: monetization. It worked, but it felt hollow. I realized this space needed to be less of a revenue stream and more of a personal record—a logbook for the trenches. So, the old system had to go.

The first casualty was the n8n automation engine. With a few focused hours of work, I ripped it out completely. It felt good. It wasn’t about building something bigger, but something with more intent. This isn’t a content farm anymore; it’s my personal archive, and it needed a system built for that specific purpose.

The New Ghost in the Machine: Python + Gemini

The overhaul replaced the broad automation of n8n with a lean, custom-built Python listener. The new workflow is far more personal and, frankly, more efficient for my needs. Now, instead of sitting down to formally write a post, I can just drop raw thoughts, notes, and images directly into a dedicated Telegram chat. It’s instant and frictionless.

From there, the Python listener grabs the data and feeds it to Gemini 3.1 Pro—the expensive, intelligent model that’s more than capable of handling the grunt work. I’m more fluent in Korean, so wrestling with English prose can be a bottleneck. This system removes that friction. Gemini takes the skeleton of my ideas and fleshes them out into coherent English. But make no mistake: the framework, the core concepts, and the final edits are all mine. It’s not an AI writing my blog; it’s an AI acting as my high-speed translator and scribe.

A Coder’s Diary, Unfiltered

As of today, this blog is officially a manual operation—my personal diary. The irony is that I’m using a sophisticated, automated backend to create a purely manual-feeling front end. It’s no longer about pumping out content. It’s about creating a detailed, unfiltered record of my thoughts and the complex automation systems I’m building day in and day out. This is where I’ll document the process, the problems, and the breakthroughs as they happen.