01 · The thesis
Retention is the default. We build the opposite.
Most software is engineered for retention. Every screen is tuned to bring you back, hold you longer, and turn your attention into revenue. That model works, and it is quietly hostile to a person in a hard season.
We build the opposite. Software for heartbreak, grief, recovery, and the other hard stretches of a life, made to be needed less over time, not more. We call the category sovereignty tech. The person owns their own return, at their own pace, and nothing here is built to keep them.
02 · The first law
No product we make is built to increase engagement.
This is the one law, and it holds across every companion and every screen. No streaks. No loss aversion. No notification engineered against a person’s own interest. No mechanic that would read as manipulative if we explained it out loud.
The story we tell a person and the story we tell an investor have to be the same story. If a feature cannot pass that test, it does not ship.
03 · The three disciplines
How the thesis becomes software.
Somatic engineering
Reach the body first.
Stress and the hard seasons live in the nervous system, below thought, before language. So we design color, pace, sound, and rhythm to reach the body before a word is read. This is somatic-first AI: the interface does the quiet regulating work before the content ever speaks.
Intelligence
A companion that stays present.
We build AI companions with agency but not authority. They draft, and the person always confirms. A companion stays present in the corner, on the visitor's side, rather than being a tool you summon and dismiss. It listens first, and it knows when to hand off to a real person.
Sovereignty
Nothing harvested, nothing sold.
Your work and your progress stay in your hands, private by default, on your own device where it can be. The most personal data never becomes the product. This is sovereignty tech made literal: the person owns their return, and the studio builds so it can never read what it does not need.
04 · The measure
A person who heals and leaves is the win.
Most studios measure success by how long they hold you. We measure it by whether you needed us less this month than last. Built to be outgrown is not a slogan, it is the metric. When a season ends and a person walks away lighter, the software did its job.
The same holds for the practice. When client work ends, the client keeps everything: the code, the accounts, the playbook. We build things to be outgrown, and that includes the studio.
05 · What we refuse
A short list, held at any price.
- HarvestWe do not collect what we do not need, and we do not sell what people trust us with.
- HookWe do not build streaks, dark patterns, or loops engineered to pull someone back against their own interest.
- HoldWe do not design for dependency. Nothing here is meant to become a habit a person cannot leave.
And a short list of work we decline outright: ad-click optimization, anything predatory toward vulnerable people, and real-money gambling. The studio builds calm, honest, protective software, or it does not take the work.
06 · The flywheel
How the work compounds.
The studio grows the honest way. Each finished product is the proof, and the proof is what brings the practice its clients. The loop feeds itself, and no product is ever asked to become a funnel into another.
- 1
Build in the open
Every product ships where people can watch it being made, doctrine and playbook included.
- 2
The craft is the proof
A finished product is the whole argument. No case study is louder than software that holds someone on a hard day.
- 3
Trust arrives unbought
Founders and teams who want software that feels human come to the studio because they have already seen the work.
- 4
The practice funds the build
Client work, not consumer subscriptions, pays for the next product. The revenue and the values point the same way.
- 5
Back to the build
Each engagement sharpens the toolkit, and the sharper toolkit makes the next build, studio or client, better and faster.
The craft moves between our products. A person never does.
The somatic engine, the companion pattern, the design system, the stack: those compound across everything we make. A person's private season never does. Routing someone's heartbreak into a pitch for their recovery would be the exact mechanic we exist to refuse.
07 · Why it holds
The stance is the defensibility.
What reads as a limit from the outside is the moat from the inside.
Three of these make the stance hard to copy. The fourth makes it hard to fake.
Conviction
The refusal to build for engagement is not a preference we could reverse next quarter. The large platforms cannot adopt it without breaking the economics they run on. For us it costs nothing, because the studio was never built to hold anyone.
Taste
One studio, one system: the typeface, the single clay accent, the gate mark, the bird companions, the somatic feel. Work made by one hand reads as intentional in a way a committee cannot manufacture, and it compounds every time the system is reused.
Model
The revenue comes from the practice, not from consumer subscriptions. The studio is not in a winner-take-all fight for anyone's attention. It sells the ability to build distinctive, humane software, which holds regardless of what the large platforms ship next.
Depth
The hardest models come from proximity, not from a market gap. Marrow's recovery does not reset on a relapse, and Surge is a clinical instrument a provider hands you. Generic companion software has nothing like them, because it was built from a category instead of from the inside.
08 · In the open
Built where you can watch.
The studio builds in the open. This page is the doctrine. The journal is written as the work happens, and every brand playbook is public in the library. Nothing about the method is hidden, because the method is the point.
The portfolio, honestly stated: nine products, four in open beta and five in development. One holds the launch lane at a time, chosen by readiness rather than a calendar. When it clears its gate, the lane passes to the next, and the outgoing product is maintained, not abandoned. Sanctuary is the flagship, and by its own doctrine it is never finished.
A product earns the lane when
- The core loop is complete and stable.
- The companion works end to end.
- The brand playbook is finalized.
- No unresolved trust or safety gap.
- At least one real person has used it start to finish.
This is what building in the open looks like from the inside: slower than a launch calendar, and much harder to fake.