Implementation:Kubeflow Pipelines Load Component From URL

Sources	Kubeflow Pipelines, KFP Components
Domains	ML_Pipelines, Component_Reuse
Last Updated	2026-02-13

Overview

Concrete tool for loading reusable pipeline components from remote YAML URLs provided by the KFP components module.

Description

kfp.components.load_component_from_url() fetches a YAML component specification from a URL and returns a callable component factory function. This is used at module level to define component operators that can be called within pipeline functions. URLs typically point to specific Git commit SHAs for reproducibility.

Usage

Call at module level to load remote components before pipeline definition. The returned factory is called within @dsl.pipeline functions.

Code Reference

Source Location: Repository: kubeflow/pipelines, File: samples/core/XGBoost/xgboost_sample.py (L4-21)

Signature:

def load_component_from_url(
    url: str,
    auth: Optional[Tuple[str, str]] = None,
) -> Callable:
    """Loads a component from a URL pointing to a component YAML file."""

Import:

from kfp import components

I/O Contract

Direction	Name	Type	Required	Description
Input	url	str	Yes	URL to component YAML file
Output	return	Callable	—	A component factory function with parameters matching the YAML spec's inputs

Usage Examples

Loading XGBoost components (xgboost_sample.py)

from kfp import components

chicago_taxi_dataset_op = components.load_component_from_url(
    'https://raw.githubusercontent.com/kubeflow/pipelines/e3337b8bdcd63636934954e592d4b32c95b49129/components/datasets/Chicago%20Taxi/component.yaml'
)
xgboost_train_on_csv_op = components.load_component_from_url(
    'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'
)
xgboost_predict_on_csv_op = components.load_component_from_url(
    'https://raw.githubusercontent.com/kubeflow/pipelines/31939086d66d633732f75300ce69eb60e9fb0269/components/XGBoost/Predict/component.yaml'
)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment