Bouletteproof

Bouletteproof builds supervised multi-agent systems for software delivery: the system around the model that decides how work is broken down, how context is fed in, and whether the output actually does what was asked before it ships. We build on one measured bet — the model is the smallest part. Most of the quality of AI-written software is the system around the model: the orchestration, the verification gate, and the learning loop that improves from its own mistakes. That's an uncomfortable claim to lead a company with, because the industry is organised around the model. We'd rather show the architecture than argue about it — and publish the results, including the ones that surprised us. If you're weighing build-versus-buy on agent infrastructure, tell us what you're working on.