Deterministic pseudonymization across multiple related CSVs — the same passphrase maps the same value to the same fake everywhere, so joins between anonymized files still work.
Tip: use samples, upload, copy, download, and send-to actions inside the workspace where available.
Referential Anonymizer replaces identifying values with fake ones deterministically: the same input plus the same passphrase always produces the same output, across every file you process. That is what makes it different from ordinary masking — anonymise a customers file and an orders file with one passphrase and the customer ids still line up, so the joins between them survive. You get a dataset that is safe to share and still usable for testing, and no key table to protect.
Re-running with the same key reproduces the exact same pseudonyms. Keep it secret — anyone with the key can regenerate the mapping.
No key set — output is still deterministic, but anyone running this tool with a blank key gets the same mapping. Add a passphrase for a private mapping.
