Predictive Maintenance MCP Server
Give any AI assistant the ability to analyze vibration data, detect machinery faults, and generate professional diagnostic reports — through natural conversation.
An open-source MCP server and Claude Code plugin that turns LLMs into condition monitoring assistants. Engineers describe what they need in plain language; the AI calls the right analysis tools and delivers results. It's designed to support and accelerate expert decision-making, not replace it — the engineer stays in control.
Quick Start
pip install predictive-maintenance-mcp
Add to your Claude Desktop config (%APPDATA%\Claude\claude_desktop_config.json on Windows, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"predictive-maintenance": {
"command": "predictive-maintenance-mcp"
}
}
}
Restart Claude Desktop. You're ready — try: "Load real_train/OuterRaceFault_1.csv and check if the bearing is healthy."
More options: install from source · VS Code setup · Docker / HTTPS deployment · use with local LLMs (Ollama)
See It in Action
Full diagnostic workflow: load signal → spectral analysis → fault detection → severity assessment → report generation
What Can It Do?
Upload a vibration signal → get a professional diagnosis through conversation.
| You say | The AI does |
|---|
| "Is this bearing healthy?" | Loads the signal, runs spectral analysis, checks for fault patterns, classifies severity |
| "Generate a full diagnostic report" | Produces an interactive HTML report with charts, fault markers, and severity assessment |
| "Extract specs from test_pump_manual.pdf and diagnose the signal" | Reads the equipment manual, looks up the bearing model, calculates expected fault frequencies, matches them against the signal |
| "Train an anomaly detector on my healthy baselines, then flag anomalies" | Trains a machine learning model on normal data, scores new signals, highlights outliers |
The AI doesn't guess — it calls 52 specialized MCP endpoints (46 tools, 2 resources, 4 prompts) running locally on your machine. Your data never leaves your infrastructure.
See the full endpoint list (52 MCP endpoints: 43 tools, 1 resource, 4 prompts)
Signal Acquisition (7 tools + 2 resources)
| Endpoint | Type | Description |
|---|
load_signal | Tool | Load vibration file (CSV, WAV, MAT, NPY, Parquet) |
list_signals | Tool | Browse available signal files with metadata |
list_stored_signals | Tool | List cached signals in memory |
get_signal_info | Tool | Signal metadata (sampling rate, duration, stats) |
generate_test_signal | Tool | Create synthetic signals for testing |
clear_signal / clear_all_signals | Tool | Cache management |
signal://list | Resource | Browse all signal files |
signal://read/{filename} | Resource | Read signal metadata |
Spectral & Statistical Analysis (10 tools)
| Tool | Description |
|---|
analyze_fft | Frequency spectrum with automatic peak detection |
analyze_envelope | Envelope analysis for bearing fault detection |
analyze_statistics | Time-domain features (RMS, kurtosis, crest factor) |
compute_power_spectral_density | Power spectral density (Welch method) |
compute_spectrogram_stft | Time-frequency spectrogram |
extract_features_from_signal | 17+ statistical and spectral features |
compute_envelope_spectrum_tool | Envelope spectrum computation |
plot_signal / plot_spectrum / plot_envelope | Visualization tools (3 tools) |
Diagnostics & Health Assessment (14 tools)
LGDiMaggio