Skip to content

Featrix Documentation

Private foundational models for structured learning.

Featrix builds custom analytics foundations trained directly on your data — reducing risk, improving accuracy, and collapsing the cost of AI.

Quick Start

from featrixsphere.api import FeatrixSphere

featrix = FeatrixSphere()

# Train a Foundational Model on your data
fm = featrix.create_foundational_model(data_file="your_data.csv")
fm.wait_for_training()

# Create a predictor
predictor = fm.create_binary_classifier(target_column="your_target")
predictor.wait_for_training()

# Make predictions
result = predictor.predict({"column1": "value1", "column2": 123})
print(result.predicted_class, result.confidence)

Installation

pip install featrixsphere

Documentation Sections

Section What You'll Learn
Why Featrix How Featrix solves every problem that makes traditional ML fail
User Guide Step-by-step guides for training models and making predictions
Technical Deep Dive Under the hood: architecture, safety, and what makes Featrix different
Use Cases Complete implementations for common ML tasks
API Reference Complete API documentation for developers and agents

What Makes Featrix Different

No hyperparameters. Your data is the configuration. No learning rates, batch sizes, or architectures to tune.

No feature engineering. Self-supervised learning discovers relationships automatically.

No silent failures. Guardrails warn you when predictions are unreliable. Calibrated probabilities you can trust.

Train once, predict many. The Foundational Model learns your data's structure. Spin up new predictors in minutes.

Support