We are building the system we kept wishing existed.
Regnum Personia started with a simple, annoying observation: the information needed to run a healthcare facility well is already being collected. It is just scattered across files, heads and habits where nobody can act on it in time.
about-hero.jpg · 2100×900
It began with a stock count that did not balance
Every practice has a version of this afternoon. The count is off. Somebody remembers a delivery that was never captured. A box of something expensive is found behind a box of something cheap, three weeks past its date. Everyone agrees this should not happen again, and then the week resumes.
We kept meeting people who were excellent at their jobs and badly served by their tools: nurses doing arithmetic on the back of a form, practice managers rebuilding the same pivot table every month, owners guessing at whether a site was profitable.
Regnum Personia is the answer we wanted to hand them: one place where stock, money and demand sit together, close enough to notice when one of them starts drifting.
about-story.jpg · 1400×1050
Give every healthcare facility, however small, the kind of operational foresight that until now only large groups with a data team could afford, and make it cheap enough and simple enough that using it is the easy choice.
Near term
Prove it in one working healthcare facility. Take their existing records untouched, measure the starting position honestly, and produce a forecast accuracy figure that survives being questioned. The pilot.
Medium term
Groups run several sites side by side, comparing consumption and spend, and moving stock between them before anything expires.
Long term
Regional pictures of shortage and demand built from consenting practices, so supply problems are seen coming rather than reported after the fact.
Seven principles we build against
These are not posters on a wall. Each one has cost us a feature at some point.
Regnum runs on your own machine or your own server. Patient-adjacent and financial records do not need to take a trip to somebody else's data centre for a stock report to work. If you later want multi-site sync, you choose it deliberately. It is never the default.
Every forecast, risk score and anomaly flag shows what it looked at and how confident it is. Staff should be able to disagree with the software and be right, which means the software has to make its reasoning inspectable.
Nobody is going to retype three years of history. Regnum imports the messy spreadsheet as it is, tells you which rows it could not read, and lets you fix them in place rather than demanding a clean file up front.
An alert that fires every day is wallpaper within a week. We would rather send three notifications a month that all deserve a response than thirty that teach people to ignore the badge.
Before Regnum is allowed to predict anything at a site, we record what that site's position actually is: counted stock against recorded stock, stock-outs, expiry, waste, emergency orders, lead times. It is slower and less impressive than a demo, and it is the only way an improvement six months later can be shown rather than asserted. When the forecasts are wrong, the error stays in the ledger.
When consumption stops looking like itself, Regnum says so and stops there. It does not say theft, and it does not say negligence. An unusual day is far more often an outreach clinic or a corrected error. Software that accuses people gets switched off, and deserves to be. The system points; a person decides.
The practices that most need better forecasting are usually the ones with the least room in the budget. Running locally, on ordinary hardware, is how we keep the cost of entry near zero, and it is why we will not casually add a dependency that turns into a monthly bill.
Want to shape what gets built next?
Early-access practices get a direct line to the roadmap. If something in your week is broken and we can fix it in software, we want to hear about it.