AI that listens to machines and structures before they fail
Lyttwise turns raw vibration data into clear diagnoses, early warnings and remaining-useful-life estimates, delivered as an engine that plugs into the maintenance platforms teams already use.
Failures announce themselves. Most systems miss the warning.
Bearings, gears, shafts and structures change how they vibrate long before they break. The signal is there, but it is buried in noise, spread across hundreds of sensors, and read by too few specialists. Most teams find out when something stops.
Who it is for
Platform partners
Maintenance and monitoring software companies who want predictive diagnostics inside their product.
We take vibration data from the sensors and systems already in place, whatever the hardware.
02 / 03
Understand
Physics-based analysis and machine learning work together to identify what is changing, which fault is developing, and how fast.
03 / 03
Act in time
Clear findings, a severity level and a remaining-useful-life estimate with honest confidence, sent straight into the tools your team already uses.
03Platform
What the engine does
01
Hybrid diagnostics
Classical vibration analysis aligned with ISO 13373, combined with machine learning. Explainable findings, not black-box scores.
02
Fault classification
Identifies common rotating-machinery faults such as imbalance, misalignment, looseness, and bearing and gear defects.
03
Remaining useful life
Tracks degradation over time and estimates how long an asset can safely keep running.
04
Calibrated confidence
Every estimate comes with an uncertainty range, so teams know when to act and when to keep watching.
05
Built for many assets
Designed from day one for fleets: many sites, many machines, many customers, with each kept separate and secure.
06
Integration-first
A documented API and webhooks, so results land inside CMMS, EAM and monitoring platforms rather than another dashboard.
04For platforms
The diagnostic engine behind your platform
If you build maintenance, asset-management or monitoring software, Lyttwise adds predictive diagnostics to your product without building a vibration-analysis team.
01Your brand, your customers, your interface. We stay in the background
02Simple API and webhooks; results map to your assets and work orders
03Pricing designed for partners, aligned with how you sell
04Pilots to prove value on real machines before you commit
05Technology and research
Built on research
Lyttwise grows out of years of research in probabilistic deep learning, signal processing and vibration-based monitoring of machines and structures. We are now working toward a foundation model for vibration data: one model that learns the physics of how things vibrate and adapts quickly to new asset types with little labelled data.
Fig. 04 · Spectrum, illustrativeamplitude / frequency
Selected publications
Selected publications coming soon.[PLACEHOLDER: 3–6 selected papers with links]
06About
Who we are
Lyttwise is founded by two engineers who have spent their careers on how machines and structures vibrate, and how to read what that vibration means.
ZS
Dr M. Zohaib Sarwar
Co-founder and CEO · Trondheim, Norway
PhD, MTech
Researcher at NTNU working on AI for condition monitoring of structures and machines; PhD in probabilistic deep learning and signal processing.