Numica is a calculation site for Italian accountants, payroll advisers and for anyone who has to answer,
quickly, a question with a number inside it: what is left in a payslip once you start from the gross annual
salary, what an employee really costs the company, what a late tax payment costs once the reduced penalty and
the statutory interest are added, whether the flat-rate regime beats the ordinary one, and what a set of financial statements says once the early-warning ratios and the DSCR are computed. It is
not a management system, it asks for no account, it keeps no files.
What sets it apart is a matter of method rather than of function. A tax calculator receives data that describe a
person or a company: income, children, a home, what is owed to banks and to the revenue. Numica is built so that
those data leave no trace. No third-party service loads into the page, no profiling cookie, no
log of the inputs. The request comes in, the answer goes out, and nothing is written in between. The
rest follows the same rule: every calculator shows its own formulas, the legal sources and the parameters for
the year, and every parameter declares the circular or the decree it comes from.
Who it is for. Three recurring uses, and they differ. The professional who has to answer while
the client is still on the phone, and cannot open the practice management system for an estimate. The business
owner or private individual checking a job offer, a payslip or a rental proposal without booking an appointment
for a three-minute question. And whoever is preparing a decision — which hiring incentive to use, whether to
distribute profits or leave them in the company, whether to buy a car or lease it — and needs the two scenarios
side by side, with the numbers that separate them.
Why it sits in this catalogue. Numica contains no artificial intelligence, and does not pretend
otherwise: it is statutory arithmetic, verified case by case. It belongs here because it shows the part of the
craft that stays the same when the material changes, the discipline that also holds up the systems where a model
really does decide: parameters in one place, formulas in the open, results put to the test before every release,
and a data practice that can be described in full on a single page.