About Us
We're building the compiler-truth layer for AI coding agents.
cat origin.log
Project Background
Nudox started as a hackathon project in November 2025. The original idea was a better documentation site for individual codebases — but the more we built, the more it turned into something bigger.
What we landed on was a way to ground AI coding agents in the actual truth of a codebase, not what a model assumes it should be. Our focus now is cutting down on hallucinations when AI agents work inside large, enterprise-scale codebases.
Current state of the world
The limits of model capability
Exponentially Increasing Code Generation
Open-source activity, by year
Mission
Project Mission
We want coding agents to genuinely understand codebases as they evolve.
The amount of code in the world is growing relentlessly, and agents cannot keep up. They miss the newest improvements to a codebase, and they cannot hold an enterprise-scale repository in memory. We are building an efficient solution that lets coding agents understand a codebase as it actually is today - not the snapshot they were trained on.
Project Goal
We want AI agents to stop guessing. Every suggestion an agent makes should be grounded in what the compiler actually knows to be true — not a plausible-sounding hallucination.
-
Cut down on hallucinated APIs in AI-generated code.
-
Make compiler-truth queryable in milliseconds.
-
Support every major language and package ecosystem.
Meet the Team
Meet the founders.
Miles Wirht
Founder & CEO
Developer with 5,000+ open source contributions. Leads development of the compiler and core infrastructure.
Robert McCrary
Co-Founder & CPO
Founded two startups before Nudox. Leads MCP development and customer integrations.
Lindsey Ferguson
Co-Founder & CTO
Microsoft engineer with 5+ years of professional development experience. Leads database design and development.
Mikita Slabysh
Co-Founder & CFO
CDW engineer with prior startup experience. Leads frontend development and customer outreach.