Machine vision, PLC tooling and command-line utilities that run on your own machines and show you why something happened. They live in git and work with the AI agent you already use.
[ 01 ]HOW WE BUILD
Yours
Runs on your own machines, and works with the AI provider you already have a contract with.
Reproducible
Every change is a commit. What runs on the line is what is in the repository.
Explainable
Structured errors, replayable inputs, and tools that show why something happened.
Agent-ready
Built so AI agents can read, build and test alongside the engineer.
Four tools
cernex
Industrial machine vision where the configuration lives in git and debugging comes first.
Opentwincat-mcp
Lets AI coding agents build, configure and debug TwinCAT 3 PLC projects.
Opencopia-cli
Copia on the command line: repositories, issues and pull requests next to git.
Openagent skills
Open skills for coding agents: a strict git workflow and structural wireframes.
OpenBuilt to order.
Beyond our own products, we design and build software on commission. We like containers best, from a single app to a full stack or Kubernetes, because what runs is exactly what the files describe.
Web development
Frontend and backend, from a single page to a full application.
Desktop applications
Desktop software that feels native on each platform.
AI tooling
Skills, tools and MCP servers that let AI agents work in your systems.
Works where you already work.
Everything starts on the command line, so it can be scripted, and a person at a terminal uses the same commands as an agent. The CLI, the agent skills and inboxcraft are open source; the industrial products are paid. Every released product is documented.
github.com/qubernetic ↗