About

Arbitrary Systems builds focused software with a calm, deliberate point of view.

Arbitrary Systems is a software company building products for domains where operational detail matters and the software should feel clearer than the work it is helping organize.

The company

The work is guided by a preference for strong operational models, restrained design, and practical utility. Rather than chasing novelty for its own sake, Arbitrary Systems approaches product development as a long-term exercise in clarity, reliability, and fit.

Current product work includes The Registry, a collector-first system for wine, spirits, beer, and cigars, PranaLogic, a studio operating system for yoga and boutique fitness businesses, Group Pours, a collaborative tasting app built around GPAT, and I'm open 2, a private pulse-based planning app for small real-world plans.

What feels different

  • Products shaped by concrete roles, records, and workflows instead of abstract SaaS conventions.
  • A preference for calm, legible interfaces that still preserve operational depth.
  • Careful treatment of private, high-trust data in collector, operational, and social contexts.

Model the work honestly

Good software should reflect the real structure of the job, whether that means bottle state and storage context, a live tasting workflow, waitlists and waivers, alliance stat comparisons, or the social shape of making plans with trusted people.

Keep private details private

High-trust products need clear boundaries around what belongs to the shared system and what should remain personal, local, circle-scoped, alliance-scoped, or operator-only.

Reduce routine friction

The products should remove friction from routine work and coordination so that collectors, studio teams, players, and everyday users spend less time fighting the software around the task.

Prefer reliability over theater

Trust is earned through predictable workflows, accurate records, and careful defaults, not through louder interfaces or performative complexity.

Make agents useful, not decorative

When AI agents enter the product surface, they should help with work the system is uniquely positioned to do: cleanup, preparation, reporting, review, and safe execution against the user's real data.