UUID/GUID Generator

UUID Version Selection
Generated UUIDs
UUID Decoder & Validator
UUID Version Info
Version 4 - Random

Randomly generated UUID with 122 bits of randomness. Most commonly used version for general purposes.

UUID Structure

Standard UUID format:

xxxxxxxx-xxxx-Mxxx-Nxxx-xxxxxxxxxxxx
  • M: Version digit (1-7)
  • N: Variant bits (8, 9, A, or B)
  • Total: 128 bits / 32 hex digits
Common Use Cases
  • 🗄️ Database Primary Keys
  • 📡 API Resource IDs
  • 🔗 Unique Identifiers
  • 📊 Distributed Systems
  • 🔄 Message IDs
Quick Actions

UUID and GUID Generation, Every Version

Generate any UUID version (1-7) plus Microsoft-style GUIDs, with validation and decoding. In practice the choice collapses to two: v4 when you need a random unique ID, v7 when the IDs land in a database index and should sort by creation time. The full version-by-version breakdown lives in UUID versions explained - and if you are wondering whether a GUID is something different (mostly it is not), see UUID vs GUID.

UUID Version Overview

Version 1 - Time-based

Time-based UUID using MAC address. Chronological ordering, embedded timestamp.

Exposes MAC
Version 2 - DCE Security

UUID with embedded POSIX UID/GID. For legacy DCE systems, rarely used.

Legacy
Version 3 - MD5 Name-based

Deterministic UUID from namespace and name. Same input = same output.

Deterministic
Version 4 - Random

Random UUID with 122 bits of entropy. Most popular choice for general use.

Recommended
Version 5 - SHA-1 Name-based

Like v3, but uses SHA-1. Preferred for new name-based UUID projects.

Deterministic
Version 6 - Reordered Time

Reordered v1 for lexicographic sorting. Better for database indexes.

Sortable
Version 7 - Unix Epoch

Unix timestamp + random. Sortable, privacy-friendly, modern standard.

Recommended for time-ordered
GUID - Microsoft Format

Microsoft format. Uppercase, optional braces. For .NET, SQL Server, Azure.

Microsoft

Which Version to Choose?

Use Case Recommendation
General-purpose unique IDs UUID v4 - random, simple
Time-sortable IDs UUID v7 - better DB indexing
Deterministic IDs from names UUID v5 - hash namespace + name
Extract timestamp v1, v6, v7
Microsoft ecosystem GUID - .NET, SQL Server

UUID Questions

It's a 128-bit ID designed to be unique without needing a central server to hand them out. Database primary keys, API resource IDs, session tokens - anywhere you need "this will never collide" confidence.

v1 = timestamp + MAC address (reveals machine info). v3/v5 = hash a name to always get the same ID. v4 = pure random (most common). v6/v7 = timestamp-based but sortable. Pick based on whether you need randomness, determinism, or time-ordering.

v4 for 90% of cases - it's random, simple, and universally supported. v7 if you want IDs that sort chronologically (helps with database indexing). v5 if you need the same input to always produce the same UUID.

With v4, you'd need to generate about 103 trillion UUIDs to hit a 50% collision chance. For perspective, that's more than generating 1 billion UUIDs per second for 3 years. So... no.

Only from v1, v6, or v7 - those embed time data. v4 is pure random, so there's nothing to extract. The decoder here handles time-based versions automatically.