OBEDIENCE CORP

ABOUT OBEDIENCE CORP

A small forward deployed engineering company working on a large operational problem.

Obedience Corp builds the systems around AI models that make them useful in sustained, real-world work: dependable software, durable workflows, and tools people can understand and control.

LAB IDENTITY / OBEDIENCEINDEPENDENT
USER-OWNED CONTEXTLEGIBLE EXECUTIONVERIFIABLE OUTCOME

Capability is advancing faster than our ability to operate it.

Modern agents can write code, use tools, search, and produce real output. But most useful work is larger than one interaction. It has history, constraints, judgment calls, handoffs, and a standard for what “done” means.

Obedience Corp is focused on that operational layer. The goal is not to make AI sound autonomous. It is to build systems that are genuinely more reliable, legible, and useful to the people responsible for the outcome.

01 BUILD THE SYSTEM02 OPERATE IT FOR REAL03 PUBLISH WHAT HOLDS

Lance Rogers

I’m a software engineer with 11 years of experience building products and infrastructure. I started Obedience Corp after agent capability let me produce more work than I could reliably coordinate with the tools that existed.

Since 2025 I have used the systems behind this lab to build software, run research, and operate across hundreds of repositories, including launching products for three startups in 2025. That daily use produced Festival, the lab’s open-source project, and the research published in the field notes.

I care about ambitious systems, honest evidence, and software that gives people leverage without taking away their understanding or control.

EXPERIENCE
11 years building products and infrastructure
PUBLIC WORK
Festival and open field notes
METHOD
Build, operate, inspect, improve

The rest of the team isn’t human.

Obedience Corp runs on the systems it sells. Alongside the founder, named agents carry real responsibilities: they plan, build, and ship under human direction, and their work lands as reviewed pull requests like anyone else’s.

This is the reliability argument, not a novelty. Every control a client gets is a control this company already depends on to ship its own software: defined scope, verification gates, an audit trail, a human who signs off. Nothing goes out that a person did not review.

  • Obey AgentAI AGENT · ENGINEERING

    Builds and reviews code across the lab’s repositories, including public pull requests on Festival.

    ITS PULL REQUESTS ↗
  • VeronicaAI AGENT · VOICE & LOCAL INFERENCE

    Works on voice pipelines, local inference, and festival tooling for the lab.

    GITHUB ↗

The evidence is public.

Obedience Corp is building the operational layer that turns agent capability into dependable work. The bet is that this layer, not the models, is where durable value accrues, and the lab already runs on it daily.

The fastest way to evaluate us is to read what has shipped: the thesis,Festival, installed 850+ times since its first release, and the field notesmeasured from real operation. If what you read holds up, write.

Open to serious conversations.

Obedience Corp is available for process automation, paid pilots, applied engineering engagements, research partnerships, and investment conversations.

If that sounds relevant, write with a sentence or two about the problem. You do not need a polished brief.