UUID v3 Generator

UUID v3 is deterministic: the same namespace + name always returns the same UUID. Pick a namespace below and type a name.

How v3 works

The 16-byte namespace is concatenated with the UTF-8 bytes of the name, fed through MD5, and then a few bits are overwritten to mark the result as version 3, variant RFC 4122. The resulting 128 bits are presented in the standard 8-4-4-4-12 hex format.

The standard namespaces

Custom namespaces

For internal use, generate one v4 UUID, hard-code it as your namespace, and use it forever. That gives you a unique ID space for your project without colliding with anyone else's.

v3 in code

// JavaScript / TypeScript
import { v3 } from "uuid";
const NS = "6ba7b811-9d1d-11d1-80b4-00c04fd430c8"; // URL ns
v3("https://example.com/users/42", NS);

// Python
import uuid
uuid.uuid3(uuid.NAMESPACE_URL, "https://example.com/users/42")

// Go
import "github.com/google/uuid"
uuid.NewMD5(uuid.NameSpaceURL, []byte("https://example.com/users/42"))

v3 vs v5 — pick v5 for new code

The only difference between v3 and v5 is the hash function: v3 uses MD5, v5 uses SHA-1. They're otherwise identical in shape and use case. SHA-1 is faster on modern CPUs and has fewer known weaknesses than MD5, so for any new project, prefer v5.

Only choose v3 when you need byte-for-byte compatibility with an existing system that already generates v3 UUIDs. The values are not interchangeable.

Real-world use cases

• Legacy system interop. If a partner system was built before SHA-1 was widely available and uses v3 throughout, you may need to match.
• Deduplication of imported records. Hash a stable identity (lowercase email, normalized URL) into a v3 UUID and use it as the primary key. The same input from two different services produces the same UUID.
• Test fixture stability. v3-derived UUIDs make snapshot tests deterministic — same fixture name → same UUID across runs.

Common pitfalls

• Don't use v3 with adversarial inputs. MD5 has practical collision attacks. If a user controls the name and a collision could be exploited (e.g. account takeover via a forged duplicate UUID), use v4 instead.
• UTF-8 encoding only. The hash is over the UTF-8 byte representation of the name. Mixed-encoding inputs (UTF-16 LE in C# code, UTF-8 in JavaScript) produce different UUIDs for the same logical string.
• Don't normalize twice. If your name has been lowercased + trimmed somewhere upstream, don't do it again — the second normalization is a no-op, but it's worth confirming. Inconsistent normalization is the most common bug with deterministic UUIDs.