v1 · Shipping
Live monitoring and alarms.
All NMEA-2000 parameters, Victron readings via Bluetooth, bilge state, anchor alarm. Alarms via the Automation rules engine. Auto-discovery and device inventory.
Platform · Systems Monitoring
All the systems, on one screen.
Tanks, batteries, bilge, engine, and every sensor connected to the bus — read live, alarmed when they go wrong, and trended over time so you catch drift before it becomes a failure.
Coverage
Whatever is on the bus, you see.
32°N is hardware-agnostic. Maretron, Victron, Yacht Devices, any sensor feeding the NMEA-2000 (National Marine Electronics Association 2000) network — they all appear in the same view. You are not limited to one brand's monitoring dashboard.
Tanks
Fluid levels.
- Fresh water — volume and fill-time estimate
- Fuel — volume, consumption rate, range at current speed
- Black and grey holding tanks
- Tank trend graphs — spot gradual leaks before they empty
Electrical
Battery banks and charging.
- Voltage, current, state of charge (SOC), and temperature per bank
- Time-to-empty and time-to-full estimates
- Charge source breakdown: solar, alternator, shore, generator
- Battery cycle count and capacity trend from supported monitors
Engine
Propulsion health.
- RPM, oil pressure, coolant temperature, transmission temp
- Engine hours — feeds the maintenance schedule automatically
- Fuel flow rate and trip consumption
- Alternator output correlated with battery charge rate
Bilge
Water intrusion.
- Bilge pump state and run-hours per 24 h
- Level sensors where fitted
- Pump run-time trending — rising run time is an early warning
Anchor
Position holding.
- Drag distance from drop point with alarm threshold
- Sustained heading change indicating swinging to a new wind
- Depth at anchor and depth change alarm
Bus traffic
Network health.
- NMEA-2000 message rate per source device
- Bus load percentage — alert before saturation
- Missing devices — alarm when an expected source stops transmitting
- Unexpected new devices — alert on first appearance
Drift detection
What the AI catches before you notice.
A single reading being off is noise. A parameter trending in the wrong direction over days or weeks is a signal. The AI agent watches the trends so I don't have to chart them manually.
Solar controller output dropped 18% over 3 weeks at similar
irradiance. Likely cause: panel fouling or shading change.
Check and clean panels.
Solar · Trend
House-bank capacity at estimated 87% of rated — 230 of 265 Ah
usable. Cycle count 312. Manufacturer spec suggests capacity
check at 300 cycles.
Battery · Cycles
NMEA-2000 bus load spiked to 78% for 4 minutes at 14:32.
Source: chartplotter sending redundant position broadcasts.
Check chartplotter NMEA output settings.
Bus · Traffic
Watermaker output 14% below 3-month average for same run duration.
Pre-filter pressure differential suggests partial blockage.
Check and replace pre-filter.
Watermaker · Output
Bilge pump run-time increased from 2 min/day to 11 min/day
over 6 days. Normal range is 0–3 min/day. Investigate source
of increased water ingress.
Bilge · Run-time How it works
Read live, alarmed on rules, trended over time.
Data comes from the bus, not from polling devices.
Every sensor reading is a bus event — timestamped, source-tagged, and persisted. The monitoring app subscribes to the events it cares about. When a device stops transmitting, the app knows immediately because the stream goes silent.
The time-series history is stored locally on the hub. Trend analysis works from local data — no cloud dependency, no subscription to historical data access.
Alarms run through the Automation rules engine.
The monitoring view shows current state. The alarm that wakes you at anchor is a rule in the Automation engine — same rules engine as everything else. "House-bank SOC below 30%" is one rule that both sounds the alarm and, if you have it configured, begins load shedding.
The rules are YAML in your config repo. Not locked inside a vendor's cloud. Not reset when you upgrade. Yours.
Auto-discovery of NMEA-2000 devices.
When you plug in a new sensor, 32°N sees it appear on the bus and adds it to the device inventory. You confirm the label — "aft bilge pump" or "generator starter battery" — and it starts appearing in the monitoring view immediately.
Missing-device alerts.
If an expected source stops transmitting — wind instrument goes silent, depth sounder drops off the bus — the platform alerts immediately. Useful underway when a sensor failure at the wrong moment matters.
Correlated multi-system views.
Solar output, battery charge rate, and load combined on one graph. Engine RPM, fuel flow, and SOG together. The monitoring app can show correlated parameters from different systems because they all come from the same event stream.
AI anomaly detection is additive.
The threshold alarms fire on current values. The AI anomaly detection fires on trends. Both are useful; they catch different things. A battery at 28V right now trips the threshold alarm. A battery whose resting voltage has dropped 0.3V over four months trips the trend alarm.
Roadmap
Where this is going.
v1.5 · Next
Load shedding and tank scheduling.
Automatic load shedding when battery drops below configured threshold. Tank-fill scheduling — target fill level plus reminder when approaching a fill point.
v2 · Later
AI anomaly detection.
Agent monitors trends and surfaces drift before it becomes a failure. Watermaker output, solar yield, battery capacity, NMEA bus load. Proposals come as maintenance queue entries — you decide whether to act.
Read about the full platform.
The demo lets you explore 32°N without any hardware. The platform page has the five-layer architecture — systems monitoring sits in the apps layer, reading from the bus layer below it.