Apps
Apps that run in your browser
Nine apps: a program planning tool, two analyzers for your own measurement files and six machine-learning labs. All processing happens on your computer. The optional Vibe Assistant in the Vibration Analyzer sends only what you choose to share.
Start here · program planning
PdM Program Setup
Before analysis comes planning: which assets to monitor, with which techniques, how many sensors and of which type.
PdM Program Setup
Build your asset list from more than 30 machine types, score criticality, and get techniques, wireless sensor counts, MEMS or piezo, mounting, Fmax, lines, record length, measurement schedules, a motor current plan and a bill of materials you can export.
PdM Program Guide
The reasoning behind the tool in seven chapters: program roadmap, the P-F curve, criticality, choosing techniques, vibration sensors and mounting, IIoT and wireless networks, rollout and KPIs.
Signal analysis · live
Analyzers for your own measurements
These analyzers compute the spectral, envelope and sideband features that ML models learn from, and apply expert diagnostic rules you can use as a baseline.
Vibration Analyzer
Drop a CSV or Excel file of acceleration or velocity. Get spectrum, envelope with automatic kurtogram band, spectrogram, cepstrum, running-speed estimation, harmonic families, sidebands, fault findings with evidence, data-quality checks and a Word report. Ask the built-in Vibe Assistant (free account) to explain findings or read chart images.
Motor Current Analyzer
Motor current signature analysis from one or three phases: rotor cage, eccentricity, misalignment, bearings, stator, driven load and supply quality, with severity scorecard, envelope, time-frequency map, families and a Word report.
Machine learning labs · live
Train and compare models in the browser
Each lab compares a machine-learning method with a classical baseline on the same data, so you can see what the model adds and where it fails.
Anomaly Detection Lab
Inject bearing wear, cooling loss, spikes, noise bursts or sensor drift into 30 days of pump data. Compare Isolation Forest, PCA, an autoencoder and a load-aware model with a fixed alarm. Load your own CSV too.
Fault Classifier
Six machine conditions, 13 vibration features, kNN, Random Forest and a neural network against expert rules. See confusion matrices, feature importance, data leakage and domain shift.
Spectrogram Image Classifier
Turn vibration into 32 × 32 spectrogram images, train a convolutional neural network with TensorFlow.js and see a saliency map of what it looked at.
RUL Predictor
Learn degradation from run-to-failure histories, predict remaining useful life with a narrowing uncertainty band, and price predictive maintenance against fixed-interval replacement.
Feature Space Explorer
Pick features and see fault classes separate or overlap in a PCA view, with Fisher scores, correlations and a quick kNN check of how much your features carry.
Pattern Classification Lab
Place points or pick a pattern and watch kNN, logistic regression, decision trees, Random Forest and a neural network draw their decision boundaries live.