Playro
In progressA desktop-first Roblox AI builder. The goal is to help someone go from an idea to a playable Roblox system faster.
SokaTech is my public lab. It is where I keep project previews, Roblox experiments, infrastructure notes, and research ideas in one place. Some work is usable now. Some is still rough. I try to label the difference clearly.
I do not want this site to read like a fake startup page. The goal is simple: show the projects, show the status, and be honest about what still needs evidence.
A desktop-first Roblox AI builder. The goal is to help someone go from an idea to a playable Roblox system faster.
A cleaner product research site. It focuses on useful notes and evidence instead of cluttered affiliate pages.
My Unraid server is where I test Docker services, local automation, and AI workflows before turning them into public writeups.
These are ideas I am testing. They stay in this bucket until I have screenshots, logs, or a demo worth showing.
Short notes about what I am learning from Roblox tooling, homelab work, local AI, and building in public.
These are active builds, not finished products. I am keeping them public because it forces me to explain what works and what still needs work.
Playro is a Roblox game builder concept. The direction is desktop-first: type what you want, generate project files, scripts, UI, and a testing path for Roblox Studio.
PureFind is a product discovery and research preview. It is meant to feel calmer and more useful than normal shopping pages.
This section is for work that is promising but not ready to be shown as complete. If it needs a screenshot, a demo, or a writeup, I say that directly.
A local automation experiment using tools like n8n, ComfyUI, Ollama, and local models. It needs a workflow screenshot and output samples before I treat it as real.
A practical project for local retail workflows. The goal is to turn messy sales exports and inventory records into clearer restocking decisions.
Tooling experiments connected to Playro and Roblox scripting. The public repo exists, but the build log still needs to be written.
Experiments around local AI, Docker services, and automation on my server. I need screenshots and logs before making bigger claims.
This is the hardware I use for testing and learning. It is not a polished cloud platform. It is a homelab that lets me experiment quickly.
| Server | Unraid / Tower homelab |
| CPU | Intel Core i7-9700K @ 3.60GHz / 8 threads |
| RAM | 46GiB available to the host environment |
| Storage | 7.5TB array / 466GB cache pool |
| Services | 17 active Docker containers / local AI / automation tools |
| Preview | preview.sokatech.xyz |
This part is intentionally simple for now. These are the writeups I want to turn into real posts as the projects get better.
How I am trying to keep the site serious without pretending the work is bigger than it is.
What I am learning while turning a rough AI builder idea into a desktop app.
Notes from running services, breaking things, and fixing them on my own server.
What is useful, what is too fragile, and what still needs proof.
Small lessons from Roblox scripting and project setup.
I am a student interested in robotics, embedded systems, AI tools, and useful software. If something here connects with what you are building, reach out.