About Exesiv

Useful software, shipped in public — AI-accelerated, human-authored.

Mission

Exesiv exists to prove that one maker, amplified by AI, can design, build, and operate a broad portfolio of quality products at a pace once reserved for whole teams — and to do it transparently, so the method itself becomes as valuable as the products.

AI has collapsed the cost of producing software, but most of the output is disposable and anonymous. The scarce, durable things are judgment, taste, and someone who stands behind the work. Exesiv keeps that human layer in charge while riding the acceleration.

What this is

A solo-founder software studio — one maker operating a portfolio of AI-augmented products (developer tooling, parametric CAD, multiplayer games, and learning tools) under one brand, one method, and one standard of taste. Not an agency, not a single-product startup: a public, AI-accelerated product house.

Software should be useful, local-first where it can be, and honest about what it does. AI is leverage, not an author: every product carries a human point of view and a person accountable for it. Prefer small, composable, self-hostable tools over sprawling platforms, and reuse a shared substrate so every improvement compounds across the whole portfolio.

Values

Useful over impressive

Ship things people actually use; novelty and polish serve utility, never replace it.

Human authorship

AI accelerates; a person designs, decides, and is accountable. Every product has an owner with taste.

Build in public

Show the process, the economics, and the failures. Transparency is the default, not a marketing afterthought.

Local-first & self-hostable

Own your data and your runtime where you can; prefer tools that work without a mandatory cloud.

Small, composable, honest

Prefer focused tools that compose over sprawling platforms; be honest about what a thing does and doesn't do.

Evidence over assertion

Done means verified — tests, builds, real usage — never merely claimed.

Operating principles

—When a product can be local-first, make it local-first.
—Prefer shipping a useful slice now over a complete platform later.
—Reuse the shared substrate (registry, harness, voice, analytics, auth) before building anything bespoke.
—If a project stops teaching or earning, pivot or retire it — no zombie products.
—Keep a human in the loop for taste and accountability, even as automation grows.
—Every product is built and operated through the same method, so the method itself stays sharp.