Skip to content
user@convertitive:~$convertitive

PDF to CSV Converter (AI-Powered)

A CSV that imports cleanly the first time — right delimiter, right number format, no manual cleanup.

CSV looks like the simplest possible table format — until you account for locale. A file produced for a US spreadsheet uses commas as the field delimiter and a dot as the decimal point; a file meant for many European tools uses a semicolon as the delimiter, because a comma is already the decimal point there (“1.204,50” instead of “1,204.50”). Get the delimiter wrong and every “amount” column silently fractures into two columns on import. Convertitive Extract reads the source PDF, detects the number locale it was originally formatted in, and always writes CSV numbers in plain machine format — dot decimal, no thousands separator — with a delimiter chosen to match how you're importing it, so the file behaves the same in Excel, Numbers, Google Sheets, or a database loader regardless of what locale the original statement was printed in.

./extract

Convert a PDF to a clean CSV

Upload a statement, invoice, or receipt. Preview the extracted rows free, then export a CSV once you sign in.

How to use

  1. Upload your PDF

    Drop a bank statement, invoice, receipt, or tabular PDF into Extract. Scanned pages are supported.

  2. Preview the extracted rows free

    Extract shows the detected table on-screen before export, so you can confirm columns split correctly.

  3. Sign in

    A free account unlocks the clean, watermark-free download. Anonymous previews are watermarked.

  4. Export to CSV

    Download a CSV with machine-format numbers (dot decimal, no thousands separator) regardless of the source document's locale — ready to import into any tool.

Frequently asked questions

Will the CSV use commas or semicolons?
Extract's CSV export uses a comma-delimited format with dot-decimal numbers by default, which is what most spreadsheet and database import tools expect. Because numbers are written in plain machine format rather than the source document's locale formatting, they parse correctly on import without a find-and-replace pass.
The source PDF used European number formatting (1.204,50) — does that survive?
No, by design. Extract detects the source number locale during extraction (that's what "numberLocale" tracks internally) and always converts amounts to plain decimal-point machine format for the CSV "value" — 1204.50, not 1.204,50 — so the file imports correctly into standard tools regardless of where the original statement came from.
Do I need an account to try this?
No. You can upload a document and preview the extracted table for free. Signing in — also free — is required only to download the watermark-free CSV.
Can it read a scanned or photographed PDF?
Yes. The extraction model reads the page as an image, so scanned statements and photographed receipts are processed the same way as native digital PDFs, with scan-quality flags on any ambiguous cell.
How are multi-line descriptions handled in CSV?
A description that wraps across two or three lines on the PDF page is joined into a single CSV field, not split into extra rows — the model reads for cell meaning, not line breaks, so a multi-line memo stays as one cell in one row.
Does a bank statement CSV include a balance check?
The reconciliation verdict (opening balance plus credits minus debits equals closing balance) is computed and shown in the Extract preview and included in the Excel export; the CSV itself contains only the row data, so run the check in the app before exporting if you need the verified figure.

About

Why CSV delimiter and decimal format aren't universal

CSV — comma-separated values — assumes the comma is free to use as a field separator. That assumption fails in locales where the comma is the decimal point instead, which is why many European tools default to semicolon-delimited exports. A converter that ignores this and always emits comma-delimited, source-locale-formatted numbers produces a file that looks fine in a text editor but silently miscounts columns the moment it's imported somewhere that expects the other convention.

Machine format vs. display format

A PDF is typeset for human eyes: "1.204,50 €" reads clearly to someone in Germany. A CSV is read by software next, and software needs one unambiguous numeric convention. Extract's pipeline separates the two — it captures the source formatting during extraction (useful context, stored per document) but always writes the "value" field in plain machine format for export, so the CSV works as an input to another program rather than as a document for a person to read.

Sources & references

Authoritative references behind the math, constants, and tables on this page. Verified by Buğra Sözeri on the dates shown and re-checked at every deploy.