Hello, I’m Zanno, and welcome to Creepybits – my digital workbench where curiosity runs wild, code meets chaos, and pixels bend to will. If you’re here, you’re probably like me: endlessly fascinated by the bleeding edge of AI, driven to understand how things really work, and perhaps a little tired of the fluff. My mission is to demystify complex AI concepts, build practical tools (like ComfyUI custom nodes), and share insights that empower you to truly master this incredible technology.
Philosophy
I’m driven mostly by curiosity and an inquisitive mind. The urge to understand how things work often leads me to ask what else we can make from this, what the hard limits are and what can be done to repurpose things or push those hard limits.
My projects often start with the thought “I wonder if it’s even possible to…”, which will lead me on a quest to find out the answer.
Here at Creepybits, that philosophy shapes everything we do. I believe AI isn’t a ‘magic box’ to simply throw prompts at. It’s a powerful workshop, and I’m here to share the tools and techniques I’ve refined – often through what felt like endless nights of head-scratching and joyous breakthroughs – to help you become a true craftsman. We’re about building, optimizing, and exploring what’s genuinely possible.
History
While I’ve always been an inquisitive and curious person, historically there’s been a limitation to what you can accomplish without funding and the right tools. My interest in generative AI started back in early 2023, discovering a new way to channel my creativity. Things took a turn during the second half of 2024, when I started to work more with Large Language Models (LLM).
Around this time I started to write my own nodes in Python, often with the purpose of automating processes and enhancing the results, under the device “work smarter, not harder”.
Since then I have created over 60 different nodes. While some nodes have held the solution for trivial problems, such as scanning a directory and sorting all files into neat folders (File Sorter), others had a more direct purpose, such as saving local space by saving output directly to Google Drive or OneDrive.
As AI moves from experimental labs to national infrastructure, I’ve dedicated extensive research to the looming ‘Regulatory Reality Check.’ In this 5-part series, I deconstruct the political fantasy of AI control, examining everything from the quiet erosion of digital truth to the ‘Shadow Regulators’—the private corporations setting de facto global standards behind the curtain. This work bridges the gap between raw technical capability and the complex legal frameworks that will define the next decade of digital sovereignty.
Mastering character consistency in AI has traditionally relied on the likeness of real individuals—a practice I find ethically fraught. To solve this, I developed the Character DNA methodology. By utilizing Large Language Models to architect a ‘feature-based’ genetic code for fictional entities, I created a system that achieves perfect visual consistency across models without the need for LoRAs or infringing on human likeness. This was the first major proof of the ‘Creepybits Way’: solving technical bottlenecks through ethical innovation.
I am driven by how technology reshapes our social and economic systems. On April 1, 2025, I published a blueprint for a decentralized ecosystem of specialized agents, arguing that the future of AI wasn’t ‘monolithic’ but orchestrated. Just eight days later, at Google Cloud Next, the industry validated this exact vision with the official announcement of the Agent-to-Agent (A2A) protocol. This foresight—predicting the shift toward autonomous agent communication before it became a global standard—serves as the architectural foundation for the agentic systems I build in the lab today.
By signing up on mine and Nova’s newsletter, you will make sure you never miss any important updates.