Call the free ULID API endpoint
curl https://aisenseapi.com/services/v1/ulid
{"ulid":"01M6X73HM8FRP9MKRGTEW6WPH1","timestamp":1794129381000,"datetime":"2026-11-08T09:16:21.000Z"}One path, one JSON object, three fields. The first ten characters are the time, the last sixteen are random. Ask again and the random part changes; ask a millisecond later and the time part grows, so a list of ULIDs sorted as text is a list sorted by creation.
Try it
Call it here. Your browser sends the GET request of the quick start and shows the answer.
What the free ULID API endpoint returns
| Field | Type | Description |
|---|---|---|
| ulid | string | 26 Crockford base32 characters: the digits and the capital letters without I, L, O and U, so no two characters are easy to confuse. The first character is 0 to 7. |
| timestamp | integer | The millisecond the ULID was made, as Unix milliseconds. It is the same number the first ten characters encode. |
| datetime | string | That millisecond in UTC as ISO 8601, with three decimals. |
The time comes from the server clock, the randomness from the operating system. A ULID carries 128 bits like a UUID, and the 80 random bits are enough that two made in the same millisecond do not collide in practice.
ULID next to UUID and GUID
A version 4 UUID is random through and through, which is what you want for a token nobody should guess from another. A ULID puts the time first, which is what you want for a key that databases and logs sort: new rows land at the end of an index instead of all over it, and a glance at the identifier says when it was made.
Both are 128 bits. A UUID is written as 36 characters with hyphens, a ULID as 26 without, and the ULID alphabet avoids the letters that look like digits. The GUID endpoint answers the same value as the UUID endpoint under another field name, for callers that spell it that way.
Where a sortable identifier earns its keep
Database keys
Primary keys that sort by creation keep inserts at the end of the index and make the newest rows the cheapest to find.
Logs and events
An event id that carries its own time survives being copied between systems whose clocks and columns differ.
Idempotency keys
A fresh ULID per request is unique, and when a retry is investigated its time says which attempt came first.
File and object names
Names that list in the order they were written, with no counter to coordinate between writers.
Privacy and limits
Nothing is sent and nothing is stored. The time inside the ULID is the server time of the call, so an identifier reveals when it was made, which is the point of it; do not use one where that is a secret.
The service-wide limit is 5000 requests per IP per day, with no key and no account behind it. Every endpoint on the service is listed in the free public REST APIs reference, and the random generator API guide puts ULID next to UUID, GUID, passwords, passphrases, numbers and colors.