FleetOSby NXTBYTE Request a pilot

Every key.Every litre.Every kilometre.

FleetOS turns your trucks and tankers into a live data network. It checks the driver at the ignition, measures fuel in the tank, reads the engine over CAN and flags anomalies before they cost you.

Request a pilot Download the product brief
  • Green: driver key, geofence and duty checks passing
  • Yellow: fuel events, refills through FDMS and idling
  • Red: anomalies such as fuel drain, detours and harsh braking

Averages across 12 FleetOS reference fleets of 42 to 380 vehicles, measured 90 days after rollout.

Five things FleetOS watches, so you don't have to.

Each one starts with hardware on the vehicle, produces data you can trust, and ends in a business event someone is accountable for. Works with AIS-140 hardware, BS-IV, V and VI vehicles, and keeps working offline.

Access control

Start and stop with the driver's own key.

No key, no crank. The vehicle will not start until the driver's personal credential, an RFID card or rugged iButton, is presented to the in-cab module. Ignition is physically interlocked through a relay on the starter line.

Every start now carries a signed driver identity, so fuel economy, harsh braking, violations and on-time departures belong to a person, not a number plate.

Credentials
RFID 13.56 MHz, iButton DS1990A, NFC
Interlock
SPST auto relay on the 30 A starter line
Latency
Under 180 ms from tap to crank
Offline
512-key local whitelist with auto-sync
Fallback
Dispatcher OTP, fully audited
  1. Tap keyDriver presents card or iButton
  2. VerifySignature, whitelist and time of day
  3. UnlockRelay closes the starter line
  4. Crank and logSigned event uploaded
What happens between the key tap and the engine starting.
Attribution changes behaviour.

When every trip is signed by the person who started it, driver performance stops being anecdotal. After 60 days of keyed attribution, the gap between the best and worst driver's fuel economy typically narrows by 22%.

Fuel monitoring

Fuel level and consumption, measured at the tank.

A capacitive probe sits inside the tank, wired to FleetOS independently of the OEM gauge. It samples every second, smooths readings with a 15-sample moving median and reconciles them against distance and engine-on time.

Consumption, refuelling and pilferage each get their own event stream, so a drain at 2 a.m. never hides inside a normal day's usage.

Sensor
Capacitive rod, 0.5% full-scale accuracy
Range
200 mm to 2,000 mm tank depth
Sampling
1 Hz with median smoothing
Calibration
20-point volumetric curve per tank
Events
Fill, drain, siphon, theft while idle
Fuel level over 24 hours for MH 14 KZ 8842 Tank level falls steadily with driving, jumps with a 95 litre refill at 06:10, and drops 42 litres at 02:15 while parked with the engine off, which is flagged as pilferage. 180 L60 L0 00:0006:0012:0018:0024:00 02:15 drain −42 Lparked, engine off 06:10 refill +95 L consumption on route
A 24-hour fuel trace for one tanker, with the refill and the night-time drain separated out.
R. Singh4.6
A. Khan4.4
P. Rao4.1
S. Yadav3.9
M. Pillai3.5
K. Das3.3
Fuel economy in km/L by driver over 30 days on the same route class. The spread from best to worst is 40%.

Telematics

BS-VI CAN or BS-IV/V sensors. One data model.

BS-VI trucks are read over J1939 or OBD-II CAN with full PGN decoding. Older BS-IV and BS-V vehicles without rich CAN get the same telemetry record from discrete sensors: inline fuel flow, magnetic RPM pickup, GPS speed and an accelerometer for harsh events.

Your dashboard doesn't care which generation of truck produced the data.

Protocols
J1939, UDS, OBD-II
PGNs
65262, 65263, 65265, 61444 and more
Sampling
10 Hz CAN, 1 Hz GPS
Edge ML
Event detection within a 120 ms budget
Uplink
MQTT over LTE Cat-M1
PGNSignalValueUnit
65262Coolant temperature87°C
65263Oil pressure412kPa
61444Engine speed1,820rpm
65276Fuel level64%
65266Fuel rate18.4L/h
64892AdBlue level58%
64777DPF soot load32%
65132Seat occupiedYes-
One decoded J1939 frame from a BS-VI truck.

Perimeter and duty control

Geofence, route, duty and theft response.

One spatial engine covers four jobs. Polygon fences mark depots, customer sites and no-go zones. Planned routes are drawn next to the actual trace, so deviation is obvious. Driver-to-vehicle mapping is enforced through key auth, and duty rules encode shifts, breaks and maximum driving hours.

When a fence, route or duty rule is broken, FleetOS escalates in tiers: SMS, then siren, then engine immobilise.

Fences
Circle, polygon, corridor, multi-layer
Routes
Map-matched with 25 m tolerance bands
Duty
Shift, break, hours of service, rest
Response
Notify, alarm, immobilise
Planned versus actual route with geofences The planned route runs from the depot geofence to the customer geofence. The actual trace follows it except for a 340 metre detour, highlighted in red. Route match is 76 percent. DepotCustomer Detour 340 m Route match 76% PlannedActual
NotifySMS and app to dispatcher
AlarmIn-cab siren, call driver
ImmobiliseStarter locked at next stop
Planned route RT-019 against the actual trace for MH 14 KZ 8842, and the theft-response ladder.

Anomaly intelligence

Alerts that connect the dots across sensors.

Raw telemetry is noisy. FleetOS correlates signals that look normal on their own but together tell a story: fuel dropping while the engine is off, a truck moving with no ignition event, or mileage 28% worse than last month on the same route.

Every alert comes with a playbook: who is told, on which channel, and what evidence to check. Adaptive thresholds keep alert fatigue down.

Engine
Rules combined with IsolationForest
Channels
SMS, email, push, WhatsApp, webhook
Dedup
10-minute rolling window, severity-weighted
Playbooks
32 built in, customisable per customer
  • Fuel drop with engine offCriticalMH 14 KZ 8842 −42 L in 14 min while parked. 2 h ago
  • Moving with no ignition eventCriticalGJ 01 TR 1120 GPS moved 0.8 km without key auth. Possibly towed. 5 h ago
  • Mileage down 22%WarningTN 09 LM 4408 Route RT-019, 3.3 km/L, was 4.2. Yesterday
  • Harsh braking clusterWarningR. Singh, 8 events in 12 min on NH-48 at km 42. Yesterday
  • DPF regeneration overdueInfoMH 14 KZ 8842 Soot at 78%, depot visit scheduled. 2 days ago
Alerts from the last seven days in a 48-truck fleet. Critical alerts fell 64% over 12 weeks as driver coaching took effect.

Flagship add-on: Fuel Dispensing Management (FDMS)

Every litre out of the bowser, accounted for.

Most fuel leaks happen where fuel enters the tank. FDMS wraps any standard in-premise pump with authentication, metering and per-vehicle policy, so every drop is tied to a driver, a vehicle and a reason.

It sits between the pump's flow sensor and its dispensing valve. Nothing flows without an authenticated session, and every session leaves a tamper-evident record: who, which vehicle, how many litres, at what flow rate, against which cap.

Flow meter
Pulse output, 100 pulses per litre
Auth
RFID, iButton, app or PIN
Valve
Solenoid, normally closed, fail-safe
Caps
Per vehicle, driver, day, pump and time window
Offline
Queues locally, syncs on reconnect
Audit
Signed, immutable ledger. ATEX Zone 1 retrofit
FDMS block diagram Driver and attendant keys authenticate at the FDMS controller. The controller opens a fail-safe solenoid valve between the pump and the nozzle, counts pulses from the flow meter, enforces policy caps and writes a signed record to the ledger, which syncs to the FleetOS cloud. Driver key Attendant key FDMS controller Auth · caps · ledger Offline queue Pump Flow meter100 pulses / L ValveSolenoid, NC Nozzle FleetOS cloudSigned ledger pulsesopen / close
How FDMS controls a standard pump.
CapLimitWindow
Per vehicle180 Lper day
Per driver240 Lper day
Per pump6,000 Lper shift
Per sessionTank volumeauto-detected
Time window05:00 to 22:00configurable
Flow rate50 L/min maxanti-spill
An example policy. The first failed check rejects the session. Dispensed litres are reconciled against the tank probe within ±2%, so short fills get flagged.
Mode A

Dual key: attendant and driver

The pump attendant and the driver must both present their keys to open a session. Two signatures on every litre, ideal for busy depots where audit rigour matters most.

Mode B

Unattended driver self-service

For captive depots with no pump operator: night shifts, remote sites, private bowsers. The driver uses their own key and caps enforce the policy, such as N litres per vehicle per day, only between 05:00 and 22:00, only at Pump 2. A camera snapshot anchors every session.

From tank sensor to operations dashboard.

Edge

Key reader, fuel probe, CAN, GPS, dashcam, FDMS, TPMS

Transport

MQTT over LTE with SMS fallback, LoRa, depot Wi-Fi

Platform

Kafka ingest, rules and ML, Timescale, S3, APIs, MCP server

Consumers

Ops web and mobile, driver app, alerts, ERP and TMS, BI and AI assistants

The rest of the platform, on the same screen.

The five core features rarely deploy alone. Operations teams add modules for safety, maintenance, load and compliance, all on the same vehicle records and permissions.

Predictive maintenance

Engine hours, vibration and fault codes turned into remaining-life estimates, scheduled for the next depot visit.

AI dashcam

Dual-lens with on-device detection of distraction, fatigue, phone use and tailgating. Live 4G stream and event clips.

Tyre pressure and temperature

Per-wheel TPMS, retrofittable. Under-inflation costs 3–5% fuel; get alerts and retread scheduling.

Cold-chain reefer

Cargo temperature and humidity logs with door-open events and GPS. Evidence-grade for pharma and food SLAs.

SOS and duress

Hidden button and voice-triggered distress. Silent alarm to the control room with live GPS and optional cabin audio.

Trip reports and POD

Automatic trip summaries, e-signed proof of delivery, and billable-km reconciliation with accounting.

EV telematics

Battery charge and health, pack temperature, charging sessions and regen. Mixed diesel and EV fleets in one view.

Load and axle weight

Air-suspension pressure turned into payload estimates. Avoids overload fines and tracks empty return trips.

What changes in the P&L.

Average improvement across 12 FleetOS reference deployments of 42 to 380 vehicles, 90 days after rollout.

Theft incidents−92%
Unaccounted fuel loss−84%
Idling time−41%
Unplanned breakdowns−38%
Overtime payouts−26%
Fuel cost per km−18%
Insurance premium−11%

For an 80-vehicle fleet, hardware plus SaaS breaks even in under 120 days.

Telemetry alone doesn't save money. Attribution does.

Any tracker can show you where the truck is. With FleetOS, every signal, whether a litre dispensed, a kilometre driven, a braking event or a detour, ends with an accountable person and a contractual outcome. That's why reference fleets recover their investment in under four months.

Deploy your way

Public-cloud SaaS, a single-tenant install in your own VPC, or fully on-premise. The same data model and dashboard in all three.

Fits your systems

REST and GraphQL APIs, webhooks and an MCP server for conversational analytics, with connectors for SAP TM, Oracle TMS, Zoho, Tally and SAP B1.

Regulation-ready

AIS-140 hardware, GDPR and DPDP-compliant data handling, ISO 27001 controls, and evidence-grade trails for pharma and explosives transport.

Ready to instrument your fleet?

Ask for a pilot scoping call, a hardware sample kit or an introduction to a reference customer. A typical pilot runs for 4 weeks on 10 to 20 vehicles, with full rollout options after that.

Request a pilot WhatsApp us Download the product brief
  • Pilot statement of work
  • Hardware sample kit
  • API sandbox access
  • Reference customer call

Jubin Patel

Founder, NXTBYTE