It is tempting to treat a national ID or a bank-account format as a constant — write the rule once, and you are done. But formats move. A country amends how a national identifier is structured or checksummed, revises its IBAN account format, redesigns its vehicle-plate scheme, or adds a new region code. When that happens, every generator that hard-codes the old rule quietly starts producing yesterday’s data.
So the real question is not whether your fake-data tool is correct today, but whose responsibility it is to keep it correct tomorrow.
With Faker, the format is yours to maintain
Faker gives you data lists and string helpers, so any structured, checksummed, or country-specific format you need — a national ID, a tax number, a plate — is something you assemble and validate in your own code. That means the format rule lives in your repository: your regex, your checksum routine, your list of region codes.
When the real-world rule changes, nothing tells you. You have to notice the change, find every place your code encodes the old format, rewrite it, and re-verify the checksum logic — and a subtle miss (a region code you forgot, a check-digit weight that shifted) produces data that looks right and fails only against a stricter validator later. The maintenance burden is real, recurring, and entirely on you.
With Real Fake Data, the format is ours to maintain
Here the format lives inside the generator, and the generator is maintained centrally against the current rules for all 27 countries we cover. When a country changes an identifier, an IBAN structure, or a plate scheme, we update the generator — and because you reach it over the API, your very next request returns the new-format data. There is no package to bump, no pattern in your code to patch, no risk that one service in your fleet is still emitting the old shape.
Keeping formats current is the product’s job, not a task handed back to you. You ask for a valid Polish PESEL or a French IBAN; staying correct as the definition of “valid” evolves is on our side of the line.
Why this compounds over time
One format, once, is easy to maintain by hand. The cost is in the multiplication: dozens of identifier and document formats, across 27 countries, each with its own occasional revisions. Track that yourself and it is a standing tax on your team forever; let it be served centrally and it simply stays current underneath you.
To be precise about Faker
Faker is a library, so by design the formats you build on top of it live in your code — that is not a flaw, it is what a library is. The distinction is only about where the ongoing maintenance sits: with a self-contained library it is yours, and with a centrally-served generator it is ours.