01
Choose a version first
Use the UUID v4 generator for fresh random values or the UUID v5 generator for a stable name mapping. v3/v5 always return one ID.
- An empty name is valid; multiple lines are one name.
Identifiers
UUID generatorFresh IDs or repeatable names
Choose a version; random batches run only when you explicitly generate them.
UUID generator
Name (v3 / v5)
No name is needed for random or time-based UUIDs. Choose the version and count below, then generate explicitly.
Input is processed in this browser and is not sent to a server.
Generated UUIDs
Generation settings
Random/time batch: 1–1,000. Name: up to 100,000 UTF-16 code units, one UUID. v1/v6: at most 10,000 ticks per millisecond; v7 has no same-millisecond ordering guarantee.
Choose settings, then generate.
Use the UUID generator to prepare fixture IDs, populate an import or derive the same identifier from an agreed name. Choose the identity rule before choosing how the text looks.
The default is one lowercase v4 with hyphens. Name-based versions instead produce one result from a namespace and the exact UTF-8 name, including spaces and line breaks.
Why use this UUID generator?
Choose randomness or reproducibility
The UUID generator separates fresh random IDs from deterministic v3/v5 results. Repeating a namespace and name reproduces an ID; pressing Generate in a random mode requests a new batch.
Export the complete batch
The preview starts with 50 rows and expands by 100. TXT contains one ID per line; the JSON report includes all results and, for name-based generation, the original name and namespace.
Decide what makes two IDs the same
02
Make changes, then generate
Generation is manual by default. Only v3/v5 can update automatically after 350 ms; random and time-based versions always need an explicit run.
- Any option change, including case or hyphens, makes results stale and blocks export until you run again. In manual mode, Ctrl/Cmd+Enter generates the results.
03
Check the count before copying
The bulk UUID generator accepts 1–1,000 results in random and time-based modes. Copy and downloads include rows beyond the visible preview.
- Example and Clear show a replacement preview; Apply commits it and Undo restores the previous state.
Six versions, two kinds of identity
Random v4
Create identifiers using browser cryptographic randomness. Version and variant occupy reserved bits; the remaining 122 bits are random.
Name-derived v3 and v5
Choose DNS, URL, OID, X.500 or a custom namespace UUID. v3 uses MD5; v5 uses SHA-1. Both hash the namespace bytes and name.
Time-bearing v1, v6 and v7
v1/v6 combine time with a random multicast node and clock sequence. v7 puts Unix milliseconds first, followed by 74 random bits around its reserved fields.
Choose the text format
Uppercase and lowercase, with or without hyphens, give 36- or 32-character text for a 128-bit UUID; formatting itself does not alter its bits. This workbench requires another run after format edits, which creates fresh IDs in random modes.
Exact inputs and practical limits
- Names are bytes, not cleaned labels
- Names are not trimmed or Unicode-normalized. NFC and NFD may produce different IDs. The maximum is 100,000 UTF-16 code units; unpaired surrogates are rejected.
- Namespaces need an agreed meaning
- Custom namespaces accept canonical 36-character UUID text, including Nil and Max, without requiring a particular version or variant. You handle DNS case, URL normalization and X.500 DER conventions yourself.
- Clock order has limits
- The 100 ns tick is an index distinguishing calls within a millisecond, not a measurement of clock precision. v1/v6 use up to 10,000 ticks per millisecond per session, then report an error until the clock advances. Clock rollback increments the sequence and restarts at the actual time; ordering across rollback or sessions is not guaranteed.
- A time prefix is not a counter
- v7 does not promise generation order within one millisecond or after clock rollback. v1/v6 never read your real MAC address; their generated node can still link IDs within a session.
- Identifiers are not secrets
- Secure randomness has no weak fallback. Collisions are not impossible and name-derived IDs are predictable. This page does not generate v2/v8, short IDs, passwords or security tokens, and does not decode IDs.
UUID generator questions
What are the name and namespace in v5?
The namespace is a UUID that establishes a naming domain; the name is the exact string within it. Agree on both, including normalization rules, before using the resulting ID across systems.
Does v5 always return the same result?
Yes, for identical namespace bytes and UTF-8 name bytes. A trailing space, case change or different Unicode composition can change the result. Formatting the UUID in uppercase does not change its underlying bits.
How do v3 and v5 differ?
Both are deterministic and namespace-based. v3 uses MD5 while v5 uses SHA-1; choose the version required by the protocol or system you integrate with. Neither is a password-hashing method.
Can I recover the original name from v5?
There is no general decoding operation that recovers the name. Someone who knows the namespace can still test likely names, so hashing a name does not make it confidential.
When should I use v5 instead of v4?
Use v5 when the same agreed name should map to the same ID. Use v4 for independent random IDs when you do not need that mapping. Keep uniqueness checks appropriate to your application.
Have a question about a result?
Open the UUID generator
Use the UUID generator with your chosen version, inspect the new results and export the complete batch when the settings are final.