We build our own graphics engines and our own locally-run AI tools. Not to replace the maker — to give the maker a longer reach. Here is exactly what that means, and what it doesn’t.
Every world begins with graphite, a sculpt, a script. The concept, the characters, the look — authored by a person, on paper, before a single line of code runs. If it doesn’t start by hand, it doesn’t start. Craft is the source; everything else is amplification.
Richard writes the studio’s own rendering engines and a bespoke, locally-run LLM toolchain — tuned to our art, not the industry average. Off-the-shelf tools bend your work toward what they were trained to make. Ours bend toward what Owen already drew.
The tools amplify; they never autopilot. They propose, organize, and clean up — the artist decides. Nothing ships that a person didn’t choose, frame by frame. The machine sweeps the floor so the maker can spend the whole day making.
Our models and tools run on our own hardware, on our own premises. Our art never leaves the building to become someone else’s training data, and our pipeline never breaks because a cloud service changed its terms. Private by default, owned outright, ours to keep.
Because we own the whole stack — the engine, the tools, and the art — a project never has to wait on someone else’s hand-off.
The concepts, characters, and worlds that give a project its identity — made by hand, designed to be built, and directed from first sketch to final frame.
Real-time rendering and game-engine technology written from first principles — tuned to make hand-painted art behave like a living place, and to run well on modest hardware.
We take original worlds from prototype to release — game, film, or print — with the engine and the art built by the same small team that ships the thing.
The quiet engine behind everything: a suite of on-prem LLM and graphics tools that keep thousands of hand-made assets consistent, catch continuity errors, and hand the artist options — never the final say.
An engine, art direction, a whole world made end-to-end — or a conversation about how we work. We’re listening.