Call the free GUID API endpoint
curl https://aisenseapi.com/services/v1/guid
{"guid":"ddf203bc-4b5a-4aec-9867-761bfaaa1bbe"}That is the whole interface. One path, one JSON object, one string inside it.
Ask again and the value changes. A later call returned ea2d6c7c-5706-43e8-834c-ea5be781a807. There is no request body, no query string and no header to set, so pasting the URL into a browser address bar counts as a complete test of the integration.
The response also carries Access-Control-Allow-Origin: *, so a fetch from your own page works in the browser without a proxy in front of it.
Response fields
| Field | Type | Description |
|---|---|---|
| guid | string | A version 4 identifier. 36 characters, lowercase hex, written in five hyphen-separated groups of 8-4-4-12. |
Nothing is stored on our side. The same identifier is never handed out twice, and there is no way to ask for an earlier one back. Treat the response as the only copy you will ever see.
Where the name GUID comes from
Two communities named the same idea at roughly the same time. The standards track settled on UUID, short for universally unique identifier. Microsoft shipped its own implementation and called it a globally unique identifier, which everyone shortened to GUID.
The name stuck to the whole ecosystem around it. A .NET developer works with System.Guid. A SQL Server column is declared uniqueidentifier and holds a GUID. A PowerShell script calls New-Guid. Nobody in that world says UUID, and code review comments get written about it.
So the vocabulary follows the platform, not the format. This endpoint exists to match the vocabulary.
What actually differs from the UUID endpoint
Put the two responses next to each other and the answer is short.
{"uuid":"11ce9c91-0065-49a2-9864-ab5722a05969"}
{"guid":"ddf203bc-4b5a-4aec-9867-761bfaaa1bbe"}Both are 36 characters. Both are lowercase hex in groups of 8-4-4-12. Both are version 4, generated from randomness by the same code path on the same server. The only thing a caller sees change is the JSON key.
That matters more than it sounds. A parser bound to a field name breaks on the wrong key, and a mismatch between what a service emits and what a consumer expects turns into a mapping layer nobody wanted to write. Picking the endpoint that already speaks your field name removes the layer.
Choose the UUID API endpoint when your schema says uuid. Choose the free GUID API endpoint when your schema says guid. Nothing else about the decision carries weight.
Reading the value you get back
The layout is worth recognizing on sight, because it tells you what kind of identifier arrived.
- The version digit
The third group opens with it. In
ddf203bc-4b5a-4aec-9867-761bfaaa1bbethat is the4in4aec, and it says the value came from randomness rather than from a timestamp or a network card address. - The variant bits
The fourth group opens with a digit between 8 and b. That sample starts the group with
9, andea2d6c7c-5706-43e8-834c-ea5be781a807starts it with8. Those bits mark the layout defined by RFC 4122 and its successor RFC 9562. - The remaining 122 bits
Version and variant consume six bits of the 128. The other 122 are random, which is wide enough that ordinary applications never design around collisions.
The bit layout, the collision math and the cost of using one of these as a primary key all get the longer treatment on the UUID page. It applies here without a word of change.
Common uses
.NET interop
The string parses straight into a System.Guid with no reshaping and no case conversion first.
SQL Server keys
Drop the value into a uniqueidentifier column, or into a seed script that fills one.
PowerShell and CI scripts
Fetch an identifier from a shell that has no GUID library loaded and no package to install.
Correlation IDs
Stamp a request on the way in and carry the value through every downstream call and log line.
Test fixtures
Give each run unique record names, so parallel test jobs never collide on shared infrastructure.
Privacy and limits
These identifiers are generated on our server and travel back to you over the network. That is fine for record keys, correlation IDs and test data. It is not fine for cryptographic key material, session tokens or password salts. Generate those locally, inside the process that will use them.
Nothing about the request is retained beyond ordinary rate limiting, and the response carries no trace of who asked for it. The service-wide limit is 5000 requests per IP per 24 hours, shared across the whole catalogue of free public REST APIs.
Reach for the free GUID API endpoint when you want a value without installing anything. In a hot loop, call the generator your platform already ships instead, because that costs no network at all.