UUID Generator.
Validate. Inspect. Understand.

Generate random UUID v4, Unix-time UUID v7, or name-based UUID v5. Versions 4 and 7 use the browser’s cryptographic random generator and fail if it is unavailable. Version 5 hashes the namespace bytes and exact UTF-8 name with SHA-1, as specified for v5. SHA-1 here produces deterministic identifiers; this is not password hashing, encryption or a digital signature.

The first 48 bits encode Unix milliseconds. Values generated in this tool use random remaining bits, with no sequence counter. They sort by their encoded millisecond timestamp, but order within the same millisecond is random. A batch shares one timestamp. The default uses your device clock; a custom timestamp is useful for fixtures and may be unrelated to the actual creation time.

Choose DNS, URL, OID, X.500 or supply a custom 128-bit namespace UUID. Names are hashed exactly as UTF-8, including case and spaces, with no Unicode normalization, URL normalization or DNS lowercasing. Prepare the canonical name for your own naming convention. Empty names are allowed. In one-name-per-line mode, CRLF, LF and CR separate names; blank lines mean empty names and one final line ending is ignored. The same namespace and exact name always produce the same v5 UUID.

The validator checks the text, variant and version layout. RFC variant versions 1–8 and the special Nil and Max UUIDs are accepted. Unassigned versions and other variants are identified but marked invalid for this validator. Valid structure does not prove that a random generator was secure, a timestamp is truthful, an identifier exists elsewhere or a UUID is globally unique. Repeated values in the submitted batch are reported separately.

Use canonical 8-4-4-4-12 hex, 32 compact hex digits, braces or the urn:uuid: prefix, in either letter case. One wrapper is accepted at a time. Leading and trailing whitespace is ignored for parsing; blank lines are skipped, while original input lines and source line numbers are retained in JSON. Incorrectly placed hyphens and non-hex characters are rejected. Import accepts UTF-8 text only and omits a leading UTF-8 BOM.

Version 7 exposes Unix milliseconds. Versions 1 and 6 expose their Gregorian 100-nanosecond timestamp, Unix time, clock sequence and node field. UTC display is floored to milliseconds, while exact Unix nanoseconds and the sub-millisecond remainder remain in the detail view and JSON. Node fields are not necessarily MAC addresses. Versions 2, 3, 4, 5 and 8 do not have a timestamp decoded here. Original names cannot be recovered from v3 or v5 UUIDs.

Generate or validate up to 1,000 entries. Input is limited to 256 KiB of UTF-8; each v5 name is limited to 4 KiB. Processing runs in a worker with a 5-second startup and 10-second execution deadline. Results show 50 rows per page. Copy and TXT contain valid UUIDs in the chosen format, with optional removal of duplicates. JSON and CSV include every result row. JSON also retains original input lines, source line numbers and names; CSV omits those untrusted fields. Copy an individual formatted UUID beside its row, or copy the inspected row as JSON. Inspecting and paging retain the existing download files. Changing output format does not generate new IDs.

UUID processing happens locally in your browser. No input history is saved and no UUIDs or names are uploaded by this tool. Clear removes inputs, generated results and temporary download links. Normal site analytics and feedback services may still load. UUIDs identify records; do not treat them as passwords or a substitute for access control.