My name is Craig, and I’m the creator of Ox (https://try-ox.com). Ox averts technical debt before it’s introduced into your codebase.

At IBM, I was a software engineer who built tools to help engineers understand their codebases which always leveraged static analysis to map the source and provide a baseline understanding. After many years of fighting with executives for priority to reduce known tech debt, I really wanted to make something that simply helped the client make simple fundamental progress rather than massive overhauls.

Without intervention, the explosion of AI-generated code will lead to an explosion of tech debt. The problem is cheapest to fix when it’s introduced, so Ox reviews code changes as they happen — either just before a git push via a CLI REPL session, or automatically when a pull request is opened on GitHub — and posts specific, contextual findings with concrete suggested fixes for the author or reviewer to simply accept.

Ox was designed to be simple to adopt and use, prioritizing developer speed while remaining comprehensive. Built for Python, it should work with any language (but I haven’t tested “any” yet).

Of course, the cleanest code does not necessarily guarantee the most performant code (see Shipilev’s curve). Ox is built to improve cleanliness and maintainability, but it’s also built to learn: Ox’s engine customizes itself to the repository on install, understanding your codebase’s conventions, rulesets, and exceptions right out of the box, and it continues to learn with each PR.

If you want to give it a run, I set up a generic demo app here (https://github.com/CraigRiggins/taskflow-ox-demo). For fun, we created an opt-in leaderboard for a mini-app that scans for the cleanest and dirtiest code you can check out here (https://try-ox.com/score.html)

I’d deeply appreciate any thoughts or questions; you can find me in the comments.

If you'd like to reach out for an in person demo you can reach me at craig@try-ox.com.