UUID Generator
Generate cryptographically random UUIDs (v4) in your browser. Bulk-generate, copy one, or copy all. Nothing leaves your device.
crypto.randomUUID() — RFC 4122 version 4, cryptographically random, 122 bits of entropy each.Use this free UUID generator to create cryptographically random version-4 UUIDs in your browser. Generate one at a time or up to a hundred in bulk, choose your preferred format, and copy with a click.
How to use it
No account, no upload — it all happens on your device.
UUID versions, in plain terms
There are several formats — here's when each makes sense.
| Version | How it's made | Use for |
|---|---|---|
| v1 | MAC address + timestamp | Legacy systems. Leaks creation time and machine identity. |
| v3 | MD5 hash of a name in a namespace | Reproducible IDs (same input → same UUID). Rare. |
| v4 | 122 random bits | Default. What this tool generates. Use unless you need ordering. |
| v5 | SHA-1 hash of a name in a namespace | Like v3 but with a stronger hash. Still niche. |
| v7 | Timestamp prefix + random tail | Modern choice when you want database-friendly ordering and uniqueness. Newer; not all libraries support it yet. |
For 99% of use cases — primary keys, request IDs, idempotency tokens, file names — UUID v4 is the right answer.
Where UUIDs work and where they don't
Strengths and trade-offs.
- Great for distributed systems. Two services generating UUIDs at the same time will not collide. No coordination needed.
- Great for idempotency. Client generates a UUID for each request; the server uses it as a dedup key so retries are safe.
- Bad as primary keys in big tables. Random UUIDs scatter writes across the B-tree, slowing inserts. Solutions: ULID, UUID v7, or a sequential ID + a separate UUID column.
- Not great for URLs you read aloud. 36 characters is too long to dictate. For human-facing identifiers, use a short slug.
- Don't leak them. They look opaque, but v1 UUIDs encode the MAC and timestamp of the issuing machine. v4 has no such leak; this tool generates only v4.
Collision math, for the curious
122 random bits means 2¹²² possible UUIDs ≈ 5.3 × 10³⁶. To hit a 50% chance of a duplicate, you'd need to generate about 2.7 × 10¹⁸ of them — roughly one for every grain of sand on Earth. For practical applications, treat collisions as impossible and don't bother checking.