Your files are processed entirely in this tab. Nothing is sent to a
server — this page makes no request to any third-party domain (check
it yourself in the Network tab of your browser's developer tools).
Once the engine has loaded you can disconnect from the internet and
confirm that processing still works.
We never receive names, addresses, emails, phone numbers, counterparty
names or account numbers — we hold no lookup dictionary, we do not ask
for one and we do not accept one. But we do not claim your data is
anonymous: this is pseudonymization, not anonymization. Real amounts,
dates and the number of distinct values stay in the file, and a
determined attacker could use them (matching against public registries
or leaked databases) to re-identify records. We programmatically
eliminate accidental leakage of direct identifiers; we do not promise
full anonymity.
Excel formulas are read as their current values, not as formulas.
After restore you get static numbers rather than a
recalculating workbook — if the formulas matter for your later work,
keep the original file separately.
Loading the Python engine…
Initializing…
Mask it
Tick the columns to hide, then press the button. Each column is
masked under its own name, so the same value gets the same token in
every column that carries that name — including columns with that
name on other sheets of the workbook.
You get two files. masked.zip is the one you send to
your analyst. mapping.KEEP-PRIVATE.secretmap is the key
— it maps tokens back to your real values, so anyone holding it can
undo the masking. Keep it. Never send it, and never put it in the
same archive as the masked file. Lose it and the next run will
produce different tokens.