TypeScript · MIT · runs in the browser
A table that looks right is never checked. So this one tells you when it is not — the document has to confirm the reading before the reading counts.
A PDF, a table pasted out of one, a CSV. One engine reads all three — only the ruler changes, and the reading says which one it used.
npx truecopy a-statement.pdfThe demo runs the library in your browser. Nothing you drop leaves the page.
truecopy - statement.pdf
pdf, 1 page(s), 6 row(s)
page 1 - cut at 100, 290
3 column(s), 6 row(s)
col kind filled role
0 date 83% when
1 text 100% -
2 amount 100% money
1 02/05/2026 | AMAZON | 12,40
2 03/05/2026 | LOYER MAI | 750,00
3 05/05/2026 | SNCF | 68,00
4 09/05/2026 | EDF | 91,32
5 12/05/2026 | BOULANGER | 7,90
6 ! | TOTAL DEBITS | 929,62 <- column 0 (date) is emptyThe total row is marked because it breaks what every other row does — no list of forbidden words, no header to read.
The second field
An empty warnings is not a promise that the reading is right. It says nothing looked wrong from the shape of the page — a much smaller claim, and an honest one. Every warning is computed without knowing anything about your document.
import { readTable } from 'truecopy';
const { rows, warnings } = await readTable(file);
// rows
// [['02/05/2026', 'CARTE AMAZON', '12,40'], …]
// warnings
// ['column 3 of page 2 is filled on only 8%
// of its rows - the cut may have invented it']No kindOf, no thresholds, no roles. Two lines.
One engine, three rulers
It votes on which left edges come back row after row — and that question is the same whichever ruler measured them. So a document only has to say where its fields start, and the same two lines read all of these:
| What was dropped | Where a column starts | What the reading prints |
|---|---|---|
| a PDF | the item's x, in points | cut at 100, 290 |
| a table pasted with spaces | the character the field starts at | cut at characters 7, 22 |
| a file written with a delimiter | the field's index — a CSV lines nothing up | cut on the delimiter |
| prose, or a file that quotes its fields | nowhere | every row came back whole |
That last row is the rule and not the exception. A quoted CSV field may hold the delimiter itself, so splitting on it anyway would shift every column after it — the rows come back uncut instead, because half-parsing is the failure this library exists to prevent.
Three things it does that others do not
A statement carries opening + Σ = closing. A career record announces its own total. Set the reading against what the document declares and you have the only check that depends on no layout — and the one that catches a misread before anybody acts on it.
A value you cannot point at is a value nobody can check. placesOf(page) is laid out exactly like cellsOf(page): same indices, one gives the text, the other gives the page and the coordinates.
A column full of dates is the date column, whatever its header says. A row that breaks what every other row does is a total, a balance, a footer. It reads no label, so it survives an issuer you have never seen.
Honest about the limit
It promises to tell you when it did not. Measured on a real statement whose header labels were hidden: the page-wide cut proposed twelve columns where the table had five, and the reading came back wrong. The only reason that was visible is that the rows did not add up to the balances the document declares.
npx truecopy a-statement.pdf