We started by pulling apart the spreadsheet with Seeley's team. Every formula, lookup table and custom function was traced and written into a scope document, so both sides agreed on what the numbers should be before any code was written.
The design then followed the job, not the spreadsheet. You set a project location, and the app fills in local climate conditions you can override. You choose an application type, enter your design needs, and see how each model performs side by side. You can change the quantity of units and see the effect straight away.
The maths runs on PsychroLib, an open psychrometric library, so the air calculations follow a known standard. Fan curve charts show whether a unit can handle the pressure, and the app flags inputs that fall outside a sensible range.
Seeley's team run the data themselves. Admins can add products, upload new performance tables, update climate data and pricing, and manage users and roles, all without a developer.