Cloudinary supports archiving & downloading your resources in ZIP files, based on different criteria (e.g. resource_type, folder, tag, list of public IDs, etc.). For more information see our generate_archive documentation and the complementary blog post. This option is available both via API and UI based on multiple resources selection.
Alternatively (or as a preliminary step) you can use the Admin API to list your resources and use the retrieved data for further filtering your content, and downloading based on their URLs or public IDs.
For our paid plans, we also support automatically backing up all original resources to the customer's own S3 bucket where it will be safely stored.