Skip to main content
xplainabledocs
Xplainable v1.4 · Now with Agentic Auto-TrainSee what's new

Build models you can actually explain.

Documentation for the Xplainable Python SDK and Cloud platform. Train transparent classifiers, deploy predictions in seconds, and ship decisions with the receipts to back them up.

Search docs, methods, parameters…⌘K
Popular:InstallationXClassifierDeploy a modelHELOC tutorial
Five-minute install

From pip install to first explanation in under a minute.

  • 1 Install the SDK
  • 2 Train an XClassifier on your data
  • 3 Call .explain() to inspect any prediction
Open the quickstart Read the API
quickstart.py
1234567891011121314151617
from xplainable.core.models import XClassifier
from sklearn.model_selection import train_test_split
import pandas as pd

# 1. Load any tabular dataset
df = pd.read_csv("credit.csv")
X, y = df.drop("default", axis=1), df["default"]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

# 2. Fit a transparent model
model = XClassifier(max_depth=8, min_info_gain=0.0001)
model.fit(X_train, y_train)

# 3. Predict — and inspect *why* it predicted that
preds = model.predict_proba(X_test)
model.explain()
Where to go next
Getting started
Install the SDK, sync with the Cloud and train your first transparent model in under five minutes.
Open section
Most read
Python API
The complete reference for XClassifier, XRegressor, partitioned models, preprocessing pipelines and NLP utilities.
Open section
Playground
REST API
Drive deployments, monitors and batch predictions from any stack. Authenticated with API keys, rate-limited, fully versioned.
Open section
Tutorials
End-to-end notebooks on real datasets — credit risk, churn, returns, energy output, lead scoring.
Open section
Advanced topics
Rapid refitting, custom transformers, XEvolutionaryNetwork search and partitioned multi-segment models.
Open section
Beta
Agentic Auto-Train
Tell the Xplainable agent your dataset and your goal. It will search, train, validate and deploy a model you can explain.
Open section
What people build with it
Risk teams
Credit risk scoring you can defend in front of a regulator.
Adverse-action codes generated directly from the model — no surrogate explanation gap.
HELOC tutorial
Growth / CX
Customer churn models that come with a reason for every flagged account.
Predict + explain in one call. Hand the rep the exact lever to pull.
Telco churn tutorial
Pricing & supply
Demand & price forecasts that survive a board meeting.
Partitioned regressors per segment, with global feature trees you can publish to stakeholders.
House prices tutorial
Open source on GitHub
64 stars · MIT licensed
View repo
Join the community on Discord
Ask questions, share models, get help
Open Discord