CodeVetter
Execution-backed verification for AI-written software changes — local-first and inspectable.
Higher confidence that an agent completed the requested change correctly and safely.Public directory · shared workshop
SaaS Maker is the public directory for Fleet products and the reusable packages, skills, templates, and feedback layer that help them ship.
Position on AI
I use AI as material, infrastructure, and a collaborator—but not as the identity of the work. The studio is about choosing useful problems, building the right product around them, and maintaining what earns its place.
Representative work
Four products show the range. The complete directory remains the honest inventory.
Execution-backed verification for AI-written software changes — local-first and inspectable.
Higher confidence that an agent completed the requested change correctly and safely.An experimental browser model playground: run small language models locally with visible generation metrics.
Train, evaluate, compare, and run specialist models with evidence about what worked and why.The thesis
SaaS Maker keeps the public product inventory inspectable and publishes the shared workshop used across it, without becoming the private Fleet control plane.
AI belongs inside the craft: in the capability, the research, and the way the work gets built. It should make a product more useful or make a difficult workflow possible; it is not a substitute for product judgment.
How work earns its place
Work begins with a specific friction worth owning, not a fashionable market label or a feature looking for a product.
Models are used where they create real leverage. The product still needs a reason to exist when the novelty disappears.
Real destinations, source, changelogs, and honest operating boundaries carry more weight than invented scale or broad claims.
The studio supplies standards and shared craft without flattening every product into one design, stack, or business model.
The boundary