Was the duty on this entry right?

Send us a customs entry you have already filed. We recompute the duty on every line against the tariff schedule as it stood on that line’s own entry date and origin, and return a report showing each line where the filed amount doesn’t match, each line we couldn’t price, and why.

Free. We don’t file anything with CBP and we take no fee on anything you recover.

Send an entry See one entry checked → See a whole book →

What this is built on

source	size	who counted it	what it answers
The schedule	29,860 lines	USITC, plus the Federal Register	what the rules require · 600 trade-war measures (Chapter 99 of the schedule) over 75,504 code entries
The receipts	19,850,923 lines	US Census	what was actually assessed at entry · 2017–2026, 114 months
The required-rate book	1,290,386 lines	built by us, by reading law	what the rules required on one line in one month · 16,513 codes × 10 origins
The port ledger	2,333,867 records	US Census port records	what physically shipped · 11 seaports

Three of these were counted by somebody else. One — the required-rate book — was built here, by reading the rules — it is the only one that can be wrong in a way nobody else can check, which is why every line of it carries whether it could be settled, and why the count we could not settle is published beside the count we could. They are four different measurements and none of them divides into another. How a line is priced → · See the record →

What comes back

Lines above the schedule

Filed duty higher than the schedule required at that line’s own date and origin.

Lines below the schedule

Filed duty lower. Named as bluntly as the first — a tool that only ever finds money in your favour is a sales tool, and you already know it.

Lines it could not price, and why

Each one named, with its reason: no published rate, a Chapter 98 provision, a China-origin code outside the Section 301 scope we hold. How a line is priced →

A difference is not money owed. We tested whether this gap predicts what is recoverable and it does not — the test, and why we published it.

What the record shows

We found an error in IMF data. They confirmed it in writing Two independent ledgers, compared — and the publisher amended their public FAQ. The trade war moved exactly one number China's share fell, four countries absorbed it, and everything else held to the decimal. How faithfully does the border collect its own rulebook? A small, persistent, mapped gap — and it doubled as the rulebook doubled.

All studies → — published whichever way they came out. How far the pre-registration claim holds is set out on /nulls.

Or test it first, with no file at all

What the schedule required on that line, and what importers were actually charged on it. Full lookup →

I run a freight forwarder. That is a conflict of interest and you should weigh it. What you send is read on our own server, goes to no third party, and is never sold, shared, or used to approach you about freight. Exactly what that means →