UUID Generator

A UUID is a 128-bit identifier you can generate anywhere — on a phone, a server, ten machines at once — without a central authority handing out numbers, and still be effectively certain no two will ever collide. This generator makes them one at a time or in bulk (up to 100), in the format you need: uppercase, no dashes, or wrapped in braces. It also has a validate tab that checks a UUID and tells you which version it is.

Two versions are offered because they solve different problems. v4 is pure random — the default, right for most IDs. v7 puts a timestamp at the front so the IDs sort by creation time, which matters a lot when they're database primary keys. Everything is generated locally with your browser's cryptographic random source.

v4 (random). 122 of the 128 bits are random (the other 6 are fixed version/variant markers). The collision math is the reassuring part: you'd need to generate roughly a billion UUIDs per second for about 85 years to have a 50% chance of a single collision. In practice, treat them as unique. This tool uses crypto.randomUUID(), backed by the OS cryptographic RNG — not Math.random().

v7 (time-ordered). The first 48 bits are a Unix millisecond timestamp, the rest random. Because the timestamp leads, v7 values sort chronologically as plain strings. That single property makes them far better than v4 as database primary keys: random v4 keys scatter inserts all over a B-tree index (page splits, cache misses), while v7 keys append in order, keeping the index tight and inserts fast.

Anatomy. xxxxxxxx-xxxx-Mxxx-Nxxx-xxxxxxxxxxxx — the M nibble is the version (4 or 7), and the top bits of N are the variant. The validator here reads exactly those bits to report the version.

Case and format. UUIDs are case-insensitive; A1B2 and a1b2 are the same value. Hyphens are cosmetic — some systems store the 32 hex chars without them.

v4 vs v7. Use v4 for general-purpose IDs where order doesn't matter. Use v7 when the ID is a database primary key or anything you'll sort or range-scan by time — the index-locality win is real and measurable at scale.

UUID vs auto-increment integer. Integers are smaller and naturally ordered but require a central sequence (a problem across shards or offline clients) and leak row counts/creation order to anyone who sees them. UUIDs generate anywhere and reveal nothing — at the cost of 16 bytes instead of 4–8.

UUID vs ULID / nanoid. ULID is essentially the same idea as v7 (time-sortable) with a shorter, case-insensitive text encoding. nanoid is a compact random string for URLs when you don't need the UUID format. If you specifically need the standard UUID shape (databases, RFC-based APIs), use these.

Are they secure tokens? v4 is random enough to be hard to guess, but "hard to guess" isn't the same as "designed as a secret." For session tokens or password resets, use a purpose-built secret generator, not a UUID you might also log or expose.