ULID Generator

Generate batches of random ULIDs with a current or explicit millisecond timestamp, locally in your browser.

Generate 1-1,000 nonmonotonic ULIDs with browser cryptographic randomness. Blank timestamp uses current time once for the whole batch.

Generate a batch for development data

Set the number of identifiers and choose Process input. Leave the timestamp empty to capture the current browser clock once for the entire batch. Alternatively enter an unsigned Unix millisecond timestamp. The tool generates a fresh random suffix for every identifier with crypto.getRandomValues; it never falls back to Math.random.

Try a fixed timestamp

Load example (replaces inputs) selects timestamp 1469918176385 and count 3. The three results begin with 01ARYZ6S41, but their suffixes vary each time. Copy one into ULID Inspector to verify the timestamp and UTC date. The fixed timestamp makes a reproducible prefix, not a reproducible full identifier.

Ordering and randomness

This is the random, nonmonotonic variant: IDs generated within one millisecond are not guaranteed to sort in creation order. All identifiers in one batch share a timestamp, even if generation spans more than one millisecond. Sorting the canonical uppercase strings groups them by their encoded timestamp; a custom timestamp or a clock moving backward can make later-generated IDs sort earlier.

Randomness reduces collision risk but does not prove global uniqueness. Enforce uniqueness in your storage system. These IDs expose their encoded timestamp and should not be treated as secret authorization tokens. The timestamp records the chosen value, not verified creation time.

Input and export limits

Request 1-1,000 identifiers at once. The timestamp range is 0 through 281474976710655 inclusive. Fractional values, signs, spaces, exponent notation and leading zeros are rejected. Count uses the same plain-integer convention. Current-time mode depends on the device clock.

The TXT download contains one uppercase ULID per LF-separated line without a final newline. Changing either field clears old results. Cryptographic randomness must be available in the browser; generation fails instead of substituting a weaker source. There is no server-side counter, saved sequence or synchronization between tabs.

Format reference

The canonical ULID specification defines a 48-bit millisecond time component followed by 80 random bits. Its Crockford alphabet excludes I, L, O and U. These tools use that format, rather than the padded RFC 4648 alphabet used by a general Base32 text encoder. Processing and file creation happen locally in this browser.