Price the job.
Defend the number.

Every price starts from one table: the rates your terminal actually charges and actually pays. It covers six categories, from labor and hourly or monthly equipment through tariff services, facilities and conventions.

Each rate carries up to four money columns, because terminals do not price with one number. Cost and bill are stored separately, and each is stored separately for straight time and overtime.

Rates are stored verbatim. Overtime is never derived as a multiple of straight time. Real labor agreements do not work that way, and a quote built on an inferred overtime rate is one the customer can take apart. If an overtime rate is not known, it is not stored.

Any rate can carry a review flag meaning confirm this before quoting. That flag follows the rate into every model that uses it.

A cost model describes a repeatable piece of business: a commodity, a customer, a service. It holds the drivers the whole model shares, such as annual volume, containers or railcars per year, average units stored, storage months and default annual hours.

Lines bind to those drivers instead of repeating them. Cost lines group into people, equipment, facilities and other. Revenue lines price per unit against a units driver.

Scenarios stay out of the way until you have more than one, so the common single-scenario case remains simple.

A quote you sent in March still computes the March number in September, after the rate library has moved on. Old quotes stay reproducible, so you can always show a customer exactly what you sent and what it was built from.

When a rate does change, the model tells you which lines it affects and lets you bring them forward. Nothing is rewritten behind your back.

Most pricing tools produce a number no matter what you feed them. The engine here is a pure calculation module covered by tests that reproduce a real customer spreadsheet to the cent, and rather than output a figure quietly it tells you what looks wrong.

A driver you have not set. Storage cost that revenue does not recover. A chosen price that falls below break-even. A rate someone flagged as confirm before quoting. Each surfaces on the model rather than waiting to be found in the quote.

Pricing is a ladder rather than a single figure. It shows the price at direct cost, at each markup you have configured, at the price you chose, and at the solved break-even where annual profit reaches exactly zero. Each row carries the resulting annual revenue, annual profit and margin.

The break-even row is solved, not guessed. That is often the most useful line on the page, because pricing at direct cost does not necessarily break even: if fixed annual costs are not recovered by your per unit rate, the real floor sits above direct cost and the ladder shows you exactly where.

Output is a plain text quote you can send, with every line traceable back to the rate that produced it.

Every model has a copilot that can see the whole thing: your rates, your drivers, your terminal profile and the computed math. It is scoped to your terminal by the same isolation that protects the rest of your data.

It can search the web for external context such as published tariffs or equipment rental benchmarks, and anything external is labeled and cited as external. Your own numbers come only from the model. If a value is not there, it says so rather than estimating.

Each conversation is private to you and persists per model, so you can pick a thread back up where you left it.

It does not tell you the market rate. Manifold prices from your rate library, flags rates that have drifted or need review, and cites any external figure it surfaces. It does not claim to know what the berth down the river quoted last week, and it will not invent a benchmark to fill a gap.