minikv implements an S3-compatible API, so you can use S3 clients/tools to PUT and GET objects through its HTTP endpoints—just like a real S3 server.

However, all storage is managed locally (in-memory, RocksDB, or Sled) by minikv. It does not use AWS S3 or any cloud storage as a backend; minikv itself stores your data on local disks.

So: applications can use minikv as if it were S3, but minikv stores data locally (not in S3).