We turn ambiguous ideas into shipped products
Discovery, prototyping and delivery — powered by agents, run by people who ship. From a problem that isn't yet a specification to something running in production.
Current status: build ......... 2026-08-02. stack ......... go 1.25 / postgres / svelte 5. engagements ... 3. services ...... 4. outreach ...... human-approved. status ........ accepting work
You need a studio, not a staffing line
Most engagements start with a problem nobody has written down properly yet. Turning it into a specification is part of the work, not a prerequisite for it — and then we take it through to production, including the parts that decide whether it still works in a year.
How we work →Product engineering, end to end
Backend, frontend, and the AI in between — from a problem that isn't yet a specification to something running in production.
AI that earns its place
Agents and LLM features where they beat the alternative — and a straight answer when they don't.
Backends that hold up
Go and Postgres systems built so the invariants live in the database, not in whoever remembers them.
Discovery and scoping
A short, paid engagement that turns a vague problem into a plan you could hand to anyone — including someone else.
Selected engagements
All work →Platform Financial Advisor — the whole picture, down to the receipt
A financial health platform that shows where the money actually goes, lets you follow any figure down to the transaction behind it, and projects where the trend lands.
Eight years of history in a single trend, 7,000 transactions searchable in one field, and projections that state their assumptions. Where the data is incomplete, the screen says so instead of guessing.
The platform we run ourselves
The studio's own operations platform — adaptive interviews, research, drafting and a durable job pipeline — built as one Go modular monolith and running in production.
Running in production. The architecture has since survived a complete swap of the AI framework by changing one adapter file — which is the result we actually cared about.
What you would otherwise maintain
| Capability | Work with us weeks | Build the team ~2 quarters |
|---|---|---|
| Discovery that produces a plan you could hand to anyone | included | 3 weeks |
| A data model with the invariants in the database | included | 4 weeks |
| Durable jobs that survive a restart | included | 2 weeks |
| AI behind ports, so the framework stays replaceable | included | 4 weeks |
| Keeping up with model and protocol churn | included | Ongoing |
The thing worth hiring for is not the framework choice, which was wrong and got replaced, but the seam that made being wrong cheap.
Tell us what you're stuck on
We'll tell you straight whether we can help — and if we can't, who probably can.