Creating custom packagers

MLRun's built-in packagers cover common Python types — scalars, collections, NumPy arrays, and Pandas DataFrames. But when your function produces or consumes a type that isn't built-in, the default behavior is to pickle the object with cloudpickle (or any configured pickling module of your choice). Pickle files are opaque, not human-readable, and tied to specific Python and module versions — making them fragile and hard to inspect.

A custom packager gives you full control over how your type is serialized and deserialized, producing readable, portable artifacts (PNG images, JSON configs, CSV tables) instead of pickle blobs.

Reminder: Packing applies to function outputs (return values → artifacts) and unpacking applies to function inputs (artifacts → typed Python objects).

• Custom packager tutorials overview
• Creating a custom packager - pack-only tutorial
• Creating a custom packager - round-trip (pack + unpack) tutorial
• Creating a custom packager - itemization (bundling & unbundling) tutorial