My guess is that:

1. Most projects already have ecto. Even if you aren’t using a database, its changesets, embedded schemas, and validators are useful.
2. The rest just copy in the ~50 lines of code from here ecto/lib/ecto/uuid.ex at v3.11.2 · elixir-ecto/ecto · GitHub

Someone certainly could extract that to a library but it’s also just not that many lines of code, and they aren’t gonna change.

I have also a reimplementation of elixir_uuid | Hex and uuid_utils | Hex . Not released yet, but if it can help I make a release. GitHub - hrzndhrn/ex_uuid: UUID generator according to RFC 4122

There are quite a few forks of the repo, you could look into those. Forks · zyro/elixir-uuid · GitHub

If a library hasn’t been updated in a while, maybe the author has moved on, I’ll also look at open pull requests to see if anyone has made the changes I need, and can move to that version. Pull requests · zyro/elixir-uuid · GitHub

If the changes I want haven’t been made, then I make them yay open source!

There are other options like About — Uniq v0.6.1

Yet another option (as an alternative to UUID, with the benefit of encoding a timestamp and thus being sortable by creation date): GitHub - omgnetwork/ex_ulid: Universally Unique Lexicographically Sortable Identifier (ULID) in Elixir