Examples

Every example, two ways

A journey view names the scene it belongs to; a code view shows the real call. Use the playground to run live scikit-learn examples in your browser through Pyodide.

Live execution is available

The playground runs real Python, pandas, numpy, and scikit-learn in this tab.

Run BreezeML in your browser →

Classification

Lantern court

Rank 22 classifiers, tune the winner, read an honest card.

df = datasets.iris()
model, report = breezeml.auto(df, "species", explain_decisions=True)
breezeml.classifiers.compare(df, "species")

Regression

Measured lantern

Predict a number, with honest error bounds.

df = datasets.diabetes()
model, _ = breezeml.regressors.random_forest(df, "target")
breezeml.conformal.conformal_regressor(model, calib, "target")

Clustering

Floating lanterns

Nine algorithms, from k-means to HDBSCAN.

res = breezeml.clustering.kmeans(df, n_clusters=3)
print(res["silhouette"])

Drift monitoring

Weather change

Know when production data drifts from training.

breezeml.drift.check(model, new_df)
# live GET /drift endpoint in every deployed API

Fairness

Balance bridge

Per-group metrics and a four-fifths rule verdict.

breezeml.fairness.report(model, df, sensitive="gender")

AI agents (MCP)

Agent shrine

Agents train and explain with safe defaults.

pip install breezeml[mcp]
claude mcp add breezeml -- breezeml-mcp

Full runnable notebooks and guides live in the repository. Browse the docs.