Phase 1 MVP

Predict equipment risk before downtime becomes expensive.

MaintenX combines machine failure prediction, explainability, and AI-assisted recommendations in one operator-ready interface.
Run PredictionView API Workflow
Realtime Risk Signal

Submit telemetry and receive probability-driven failure classification instantly.

AI Maintenance Assistant

Gemini-powered explanations transform model output into actionable operator guidance.

Operational Confidence

Expose prediction, recommendation, and readiness in one clean workflow.

Platform Snapshot

Core Metrics and Reliability Signals

A compact benchmark and product-readiness layer to help teams quickly understand performance, delivery flow, and AI support status.

XGBoost Accuracy

99.90%

Primary production model benchmark

Operational Readiness

3 Layers

Prediction, explanation, recommendation

AI Assist Mode

Gemini

Natural-language maintenance guidance

Inference Flow

Realtime

FastAPI-connected dashboard experience

Sensor Inputs

Prediction Form

Ready
Machine Type
Encoded product family identifier.
Air Temperature
Ambient operating temperature in Kelvin.
Process Temperature
Process temperature in Kelvin.
Rotational Speed
Machine spindle speed in RPM.
Torque
Mechanical load in Newton meters.
Tool Wear
Accumulated wear exposure in minutes.
TWF Flag
Tool wear failure indicator.
HDF Flag
Heat dissipation failure indicator.
PWF Flag
Power failure indicator.
OSF Flag
Overstrain failure indicator.
RNF Flag
Random failure indicator.
API: https://maintenx-s395.onrender.com
Decision Output

Prediction Result

Awaiting prediction
Probability

--

Risk Score

--

Status

Idle

Submit machine telemetry to generate risk probability, classification, and maintenance guidance.

AI Assistant

Maintenance Guidance

Explanation

The assistant explanation will appear here once a prediction response is received.

Recommendation

Recommended maintenance action will appear here after inference.

Trust & Explainability

Make predictions understandable, not just accurate.

MaintenX combines model inference with explanation and recommendation layers so engineering teams can review why a machine is risky and what action should happen next.

Included in this flow

Prediction API, benchmark metrics, assistant guidance, and SHAP report generation are all represented in the product workflow.

Explainability Layer

SHAP reports expose the strongest contributors behind failure classification so operators can review decisions with more confidence.

Metrics Visibility

The product communicates probability, risk percentage, and operational state in a compact executive-style decision panel.

Actionability First

Assistant recommendations turn raw model outputs into maintenance next-steps instead of leaving teams with only scores.