UUID Format Explained: The 8-4-4-4-12 Structure

The Five Groups

f81d4fae - 7dec - 11d0 - a765 - 00a0c91e6bf6
   8        4      4      4         12        hex digits
  32       16     16     16         48        bits

The groups originally matched fields of the version 1 layout (time_low, time_mid, time_hi_and_version, clock_seq, node). In newer versions the field meanings differ, but the text layout stays the same for every UUID.

The Version Digit

The first hex digit of the third group is the version:

In f81d4fae-7dec-11d0-a765-00a0c91e6bf6, the third group is 11d0, so it's version 1. How to tell a UUID's version walks through more examples, and UUID versions explained covers each one.

The Variant Digit

The first hex digit of the fourth group holds the variant in its top bits:

• 0–7: NCS backward-compatibility variant (legacy)
• 8, 9, a, b: the standard RFC 9562 variant
• c, d: Microsoft's legacy variant
• e, f: reserved for the future

Almost every UUID you'll see uses the standard variant, so the 17th character is 8, 9, a or b. In the example, the fourth group is a765: standard variant.

Case and the Canonical Form

RFC 9562 allows upper, lower or mixed case hex digits when reading UUIDs, and recommends lowercase when generating them. The canonical form is lowercase with hyphens:

f81d4fae-7dec-11d0-a765-00a0c91e6bf6

Normalising to this form before storing or comparing avoids treating the same ID as two different values.

Other Text Forms You'll Meet

All represent the same 128 bits. Removing hyphens or adding braces changes the text, not the value.

Binary Form

Stored natively, a UUID is 16 bytes. The standard byte order follows the text left to right (big-endian). Microsoft's GUID structure stores the first three groups little-endian in memory, which is why byte-level conversions between .NET and other systems sometimes scramble the first half of the value. Native database types (uuid in PostgreSQL, uniqueidentifier in SQL Server) handle this for you; trouble starts when code converts to raw bytes manually.

Special Values

Two UUIDs fall outside the version rules: the nil UUID (all zeros) and the max UUID (all f). Nil UUID and max UUID explains what they're used for.

Checking a UUID's Format

The UUID validator accepts standard, uppercase, braced, compact and URN forms, confirms each is a correctly formed 128-bit value, reports its variant and version, and shows the canonical form. It checks up to 1,000 lines at once. For time-based versions, the UUID decoder also shows the embedded timestamp. To validate in your own code, see UUID regex patterns.

Byte Layout for Each Version (Simplified)

Bit positions count from the leftmost bit of the standard text form.

Formatting in Different Languages

• JavaScript: id.toLowerCase(), and replaceAll('-', '') for compact form.
• Python: str(u), u.hex, u.urn.
• C#: g.ToString("D"), "N", "B".
• SQL Server: CONVERT(char(36), @id) returns uppercase; wrap in LOWER() for canonical form.
• PostgreSQL: id::text returns lowercase hyphenated form.

Summary

A UUID is 32 hex digits in 8-4-4-4-12 groups. The 13th digit is the version, the 17th encodes the variant, and lowercase with hyphens is the canonical form, whatever variations your systems accept.