Built by O&G engineers · ML detection · AI diagnostics

Know what your rotating equipment is thinking

RotaBrain monitors pumps, compressors, and motors in real time — detecting anomalies before they become failures, and generating plain-English AI diagnostics your maintenance team can act on immediately.

Explore live demo

No commitment. We'll reach out to discuss your equipment and data.

Built on 20+ years O&G field experienceML-detected · AI-diagnosed · Field-validatedPowered by Claude AI (Anthropic)Data never leaves your environment

Everything a maintenance engineer actually needs

Statistical ML detection on every reading. AI-generated diagnostics your team can act on.

Multi-layer anomaly detection

Three detection methods per reading: fixed-baseline deviation (distance from established normal), rolling deviation (sudden changes vs. recent behavior), and CUSUM drift (slow sustained trends that point-by-point checks miss). Four severity tiers — normal, low, medium, critical — across vibration, temperature, pressure, and motor current.

AI diagnostic narrative

Claude generates a streamed plain-English report: named failure hypothesis, degradation rate in engineering units per day, and projected time to alert and danger thresholds. Equipment-specific confidence levels flag where z-score detection has limited coverage — so you know when to call the specialist.

Multi-channel sensor timeline

Four toggleable sensor channels — vibration (mm/s), temperature (°C), pressure (bar), motor current (A) — over a 7- or 30-day window. Anomaly periods highlighted as red zones; individual anomaly events plotted as scatter points so you see exactly when deviation began.

Health score & urgency

0–100 score computed from the 7-day anomaly window: penalizes overall anomaly rate and critical-severity events separately. Urgency tiers — Monitor / Schedule / Act Now — derived from the AI report. Refreshes on every dashboard load — no manual trigger needed.

SCADA-ready data layer

Upload a CSV export from any historian — raw sensor readings are processed in memory and never written to any database or log at any layer. Only the AI diagnostic report is stored. Built on PostgreSQL for future direct historian integrations as your deployment grows.

Multi-asset dashboard

Monitor pumps, compressors, motors, and gearboxes from a single view. Asset sidebar sorts by severity — critical equipment surfaces at the top. Health score, anomaly count, and peak z-score visible at a glance.

From sensor data to AI report in minutes

Three steps. No infrastructure. No data science degree.

01

Upload your data

Export a CSV from your historian or SCADA system and upload it directly. No special integration or data science setup required.

02

Anomaly scoring

Every sensor reading is scored against its rolling baseline. Spikes, drift, and degradation patterns surface automatically — no threshold tuning needed.

03

Get your AI report

Click Run Analysis. Claude generates a plain-English diagnostic — root cause, affected tags, recommended action.

CSV format — what your file needs

ts · vibration · temperature · pressure · motor_current
TimestampISO 8601 or standard datetime — required
Vibrationmm/s — optional
Temperature°C — optional
Pressurebar — optional
Motor currentA — optional

At least one sensor column required. Column names are flexible — common aliases like temp, vib, curr are accepted. Export directly from your historian or SCADA with no reformatting.

Equipment coverage

RotaBrain complements OEM protection systems — it doesn't replace them. Coverage scales from full diagnostics on undermonitored assets to narrative context layered on top of existing signals.

Full diagnostics

Complete health monitoring, anomaly detection, and AI diagnostics

  • Centrifugal pumps
  • Screw compressors
  • Gearboxes
  • Induction motors (100 kW–1 MW)
Complementary coverage

Derived from existing sensors — no new hardware required

  • Couplings & shaft alignment
  • Lube oil systems
  • Seal systems
  • Cooling water loops
OEM narrative layer

RotaBrain adds cross-system context on top of existing OEM protection signals

  • Gas turbines
  • Large reciprocating compressors
  • Diesel engines
  • Motors >1 MW

What the AI actually says

Not "anomaly detected." A real diagnostic a maintenance engineer can act on.

AI Diagnostic Report — Pump A · PS-A-001
Critical
CRITICAL — Bearing degradation pattern detected.

Pump A shows elevated vibration at 12.4 mm/s (+3.8σ) concentrated in the 85–120 Hz band, consistent with outer race defect frequency on the non-drive end bearing. Temperature rise of +14°C above 7-day baseline confirms friction-induced heat generation at the bearing seat.

Motor current draw elevated at +1.8σ — likely secondary to increased mechanical resistance from bearing wear, not a primary electrical fault. Suction pressure nominal — pump hydraulics unaffected at this stage.

Recommended: Schedule bearing inspection within 48 hours. Do not defer.
30-day site pilot

Evaluate RotaBrain on your actual equipment

The way industrial companies approve new technology — a time-bounded pilot with a defined deliverable, not an open-ended SaaS subscription.

What's included

  • Anomaly analysis of up to 5 assets (pumps, compressors, motors)
  • Each data upload analyzed within 24 hours — daily, weekly, or on your historian's export schedule
  • Weekly summary reports with flagged deviations
  • End-of-pilot site health report (PDF) with prioritised maintenance recommendations
  • Direct access to our Maintenance and Reliability Engineer — not a support ticket queue

How the pilot runs

Week 1

Onboarding — upload your historian's CSV export (30+ days of historical data). Individual asset baselines established; first anomaly analysis delivered within 24 hours.

Weeks 2–4

You send updated data exports on your schedule — daily or weekly. Each batch analyzed within 24 hours; deviations flagged and weekly summary report emailed to your team.

Day 30

Full site health report delivered. Continue at $299/asset/month or walk away — no lock-in.

$500/ asset · 30 days

$2,000 per site (up to 5 assets)

Then $299/asset/month · Invoice-based · No credit card · No auto-renewal

Ready to see it on your equipment?

Connect your historian or drop in a CSV. We'll run the full anomaly detection and AI diagnostic pipeline on your actual data — no commitment.

Explore live demo

Built by Argeo Alecha · hayahai.com