AI Maintenance Pro

AUC 0.98 on Hitachi's MIMII dataset. Train an equipment-specific anomaly detector in minutes — just record normal sounds with your iPhone.

Reproduce the "trained ear" of a veteran technician using AI. Record the normal operating sounds of your equipment with your iPhone, and build a dedicated anomaly detection AI for that specific machine.

■ Data Security Recorded audio is processed entirely on your iPhone — nothing is ever sent to external servers. AI training and diagnosis complete on-device, so confidential factory sounds never leave your organization. Safe to use even under strict information security policies.

【How to Use — 3 Steps】

1. Register Equipment (enter name and type)
2. Record Normal Sounds (10 seconds × 3+ times, 10+ recommended)
3. Tap "Train Model" — your equipment-specific AI is ready

After training, simply record the same equipment and the AI automatically detects anomalies.

【Key Features】

■ Equipment-Specific AI Training Learn normal sound patterns for each machine and build a customized detection model on-device. A statistical ensemble (3-sigma + Mahalanobis distance) classifies results as Normal, Warning, or Critical.

■ High-Accuracy Model — Premium (Optional) Add a server-side One-Class SVM for higher accuracy. Combined with the on-device model, a 3-detector ensemble boosts detection performance.

■ AI Detailed Analysis (Optional) Powered by Google Gemini — generates root-cause estimates and action recommendations tailored to equipment types such as bearings, gears, motors, pumps, and compressors.

■ PDF Report Export Export diagnosis results as business-grade PDF reports with anomaly scores, verdicts, and recommendations.

■ Diagnosis History Management Store past results and manage them efficiently with filter, sort, and search.

■ Offline Ready On-device AI supports recording and diagnosis without an internet connection.

【Use Cases】

■ Predictive Equipment Maintenance Detect early signs of failure from operating sounds. Prevent unplanned downtime and enable scheduled maintenance.

■ Acoustic Quality Inspection Inspect products whose quality is judged by sound — automotive sunroofs, power windows, sliding doors, power seats, motor-embedded components, gearboxes. Quantify and automate production line inspection with a single iPhone. Replace subjective judgments with objective, statistically grounded data to reduce inspection variance.

■ Incoming Inspection Acoustic inspection for purchased parts or finished goods. Lot-level quality monitoring and defect detection at delivery.

【Ideal For】 • Maintenance staff at manufacturing sites • Equipment service providers • Quality control personnel • Sensory/acoustic product inspectors • Automotive parts and precision equipment inspectors • Small and mid-sized manufacturing business owners

【Supported Targets】

■ Equipment Bearings, gears, motors, pumps, compressors, and other rotating machinery.

■ Products Automotive components (sunroofs, power windows, sliding doors, power seats), appliances, power tools, motors, gearboxes, fans, actuators — any product judged by its operating sound.

*Works with any target for which you can capture normal operating audio.

【Pricing】 • Core features: Free (equipment registration, AI training, diagnosis, report export, history) • AI Detailed Analysis: In-app purchase (one-time / lifetime access)

*Diagnosis results are for reference only. Consult a qualified technician for critical equipment decisions.

【Dataset Used for AI Model Testing】 MIMII Dataset: https://zenodo.org/records/3384388 Paper: https://arxiv.org/abs/1909.09347