UUID Generator

What is a UUID and why use one?

A UUID is a 128-bit identifier written as 32 hexadecimal digits, like 550e8400-e29b-41d4-a716-446655440000. Its whole purpose is that you can create one on any machine, at any moment, without asking a central authority for permission, and still be practically certain nobody else will ever produce the same value.

That is what makes it different from an auto-incrementing number. A database sequence needs one place to hand out the next value, which becomes a bottleneck the moment you have several servers, an offline mobile client, or data merged from two systems. A UUID sidesteps all of it: the client generates the ID before anything is saved, and nothing has to coordinate.

Which version should I use?

Four choices here, and the honest answer is that most people want the first one:

Pick v4 unless you have a reason not to. Pick v7 when the identifier is going into an indexed column, because new rows then land next to each other instead of scattering across the index, which keeps inserts cheap as the table grows.

Do v7 UUIDs really sort in order?

Here they do, including inside the same millisecond, and that is worth checking on any tool you use.

The timestamp in a v7 UUID has millisecond resolution. Generate a batch in one go and every identifier shares the same millisecond, so the ordering falls back to the random bits, which is to say no ordering at all. That defeats the only reason to choose v7 over v4.

RFC 9562 solves it with a monotonic counter: the bits right after the timestamp hold a number that increments for each UUID created in the same millisecond. This tool implements it, so a batch of a thousand comes out strictly ascending. It is a small detail that decides whether the format does what it claims.

What does UUID v5 solve?

Version 5 turns a name into a UUID, and always the same one. It hashes a namespace UUID together with your text using SHA-1, so example.com in the DNS namespace produces one specific identifier, today and in five years.

That is useful whenever you need a stable ID for something that already has a natural name: a URL, a file path, a product code from a supplier's system. Instead of storing a mapping table, you recompute the UUID whenever you need it. The four standard namespaces (DNS, URL, OID and X.500) are built in, and different namespaces give different results for the same name, which is exactly the point.

How do I decode a UUID I already have?

Paste it into the Decode tab. The tool reads the version digit and the variant bits and tells you what format you are looking at, and for a version 7 UUID it extracts the embedded timestamp and shows exactly when it was created, in your local time and in UTC.

A coloured map underneath shows which part of the value is what: the timestamp, the version marker, the counter, the variant, and the random remainder. It accepts the value with or without hyphens, wrapped in braces, prefixed with urn:uuid:, in upper or lower case.

One honest limit: a v4 UUID contains nothing to extract. It is random by design, so there is no creation time and no machine information hidden in it. Any tool claiming otherwise is guessing.

Generating in bulk and exporting

Up to a thousand identifiers in a single batch, which covers seeding a test database or filling a fixture file. The list on screen shows the first two hundred so the page stays responsive; copying and downloading always include the whole batch.

Four output formats are available: plain text, JSON array, CSV with a header row, and a ready multi-row SQL INSERT. The SQL uses a placeholder table and column name, so change those to match your schema before running it.

Formatting follows whatever your platform expects. Uppercase, hyphens removed, and braces are all one click, and braces in particular is how the .NET world usually writes a GUID.

Is any of this sent to a server?

No. Every value is generated and decoded in your browser, using the built-in cryptographic random source, the same one browsers use for security keys. Nothing is uploaded, logged or stored, and the page keeps working offline once it has loaded.

That matters more than it might seem. An identifier generated by someone else's server is an identifier someone else has seen, which is the wrong property for a primary key in a private system. Here they never leave your device.