While you are testing your algorithm, you'll likely want to call the latest version of a private algorithm vs. specifying it explicitly.
Typically, when you want to call a private version of an algorithm that you've published you will specify the hash of a git commit or the semantic version (i.e. "1.0.0"). When you are iterating quickly, you may want to always call the latest privately published version rather than updating your code with a new semantic version each time you publish a change.
Thankfully, you can specify the constant version spec latestprivate in place of a semantic version to call the latest privately published version of your algorithm using any of the Algorithmia clients.
For example, in Python you would have specify the version spec as:
algo = client.algo('deeplearning/ColorfulImageColorization/latestprivate')