Telemetry is vendor-locked and noisy
Every BMS brand speaks its own dialect. Raw frames arrive irregular, unit-inconsistent, and siloed per vendor — so no one has one clean view of the fleet.
Battery-intelligence layer
Know how many charged packs each station needs — before the rush.
Rootd turns raw BMS telemetry into per-station forecasts, so more charged packs sit in the right station at the right time — protecting daily rider earnings retained after swap friction.
Works with any BMS vendorTenant-isolatedVendor-neutral battery digital twin
Charged packs availableForecast demand
0dry-station events forecast
Packs pre-positioned ahead of both peaks.
drop: e-mobility context photo · swap station
Per-station forecasts · Any-vendor BMS · Multi-tenant · Multi-country-ready
Placeholder operator marks — shown to illustrate fit, not production scale.
The problem
Swap networks are flying with instruments they can't read.
The cost lands on riders — as downtime and swap friction — and on operators and financiers as risk they can't price.
Stations run dry — or sit overstocked
Without per-station forecasts, restocking is reactive. Packs end up in the wrong place at peak hours, and riders queue or ride on to the next station.
Battery health is invisible to lenders
Financiers underwrite packs they can't see degrade. With no trusted state-of-health signal, residual value is a guess and loan books carry hidden risk.
What Rootd is — and isn't
An intelligence layer. Not another box, app, or operator.
ROOTD IS
- Cloud intelligence on top of your existing network
- A vendor-neutral battery digital twin, derived from telemetry
- Forecasting, scheduling & residual-value signals
- Read APIs and projections your apps consume
ROOTD IS NOT
- No hardware — Rootd builds and operates none
- Not a fleet operator — we don't run stations or bikes
- Not an ERP or accounting system
- Doesn't replace the BMS or the operator's swap app
Three pillars
One twin. Three decisions it makes sharper.
Operators plan inventory, riders keep earning, and financiers see residual value — all from the same normalized signal.
PILLAR 01 · OPERATORS
Station inventory forecasting
Per-station charged-pack forecasts that enable peak smoothing.
Rootd projects charged-pack supply against rider demand for each station over the next 24 hours, then turns it into a restock plan you can act on before the morning and evening peaks.
- Forecast demand per station, not per network
- Pre-position packs to smooth both daily peaks
- Fewer dry-station events; less idle overstock
FORECAST · KAWANGWARE HUBon track
TODAY · RESTOCK PLAN4 stations
- Kawangware Hubevening peak+14by 16:00
- Ngara Junctionmorning peak+8by 06:30
- Industrial Areaoverstocked−6rebalance
- Westlands Mallwithin bandholdstable
PILLAR 02 · RIDERS
Rider scheduling
Recommendations that protect a rider's earning time.
Rootd returns a suggested swap window that minimizes travel plus wait — surfaced through the operator's own app as a projection, never a separate consumer brand.
- Suggested window from live station load
- Cuts swap friction: travel + queue time
- Output the operator's app chooses to surface
GET /v1/riders/{id}/projection200 OK
SUGGESTED SWAP WINDOW
13:20 – 13:50
Kawangware Hub · 600m away
est. wait 3 min
0%earning time retained
−11minvs. unplanned swap
recommendation only · operator app decides surfacing
PILLAR 03 · FINANCIERS
Loan-performance visibility
State-of-health-backed residual value and loan-health signals.
A trusted SoH trend per pack gives lenders a residual-value confidence band instead of a guess — so loan books are underwritten on how batteries actually age.
- SoH trend per pack and per cohort
- Residual value as a confidence band, not a single figure
- Early signal on packs drifting off their curve
SoH TREND · COHORT 12 · 12 MOconfidence ±1.2%
90.4%mean SoH, cohort
residual
confidence bandderived from SoH curve
confidence bandderived from SoH curve
The platform
Any BMS in. One twin out.
Rootd normalizes every vendor's telemetry into a single vendor-neutral digital twin — connectivity, state-of-health, and an estimated state-of-charge derived from SoH and the latest reading.
NORMALIZATION
Vendor A · CANraw BMS frames
Vendor B · MQTTraw BMS frames
Vendor C · RESTraw BMS frames
VENDOR-NEUTRAL TWIN
SoHconnectivityest. SoCcycles
LIVE TWIN · TENANT: NORTHWIND
onlinestaleoffline
PackConnectivityState of healthEst. SoCLast reading
SoC is estimated from SoH + latest reading — never stored as a peer field. Offline packs hold last-known twin state.
Connectivity is a modeled state — online / stale / offline — not an error. SoC is always shown as estimated, derived from SoH.
How it works
From raw frames to a decision — in four steps.
01
Ingest
Pull BMS telemetry from any vendor — CAN, MQTT, REST — vendor-neutral by design.
02
Normalize
Resolve units and gaps into one vendor-neutral battery digital twin.
03
Forecast & recommend
Project station inventory, swap windows, and SoH-backed residual value.
04
Decide
Operators restock, riders swap, financiers assess — on the same signal.
Trust & security
Built multi-tenant from the first query.
Confidence comes from boundaries you can verify — isolation, observability, and honest data states.
Tenant-scopedAudited readsStale-state modeled
Tenant isolation
Every query is tenant-scoped and bounded — one operator's data is never reachable from another's.
Observability
Ingestion, freshness, and twin state are monitored end-to-end, with audit trails on every read.
Data integrity
Connectivity loss is modeled as stale or offline, not silently dropped — the twin holds last-known state.
FAQ
Straight answers about scope.
No. Rootd is an intelligence layer on top of your existing BMS. It reads telemetry the BMS already produces and builds a digital twin from it — the BMS keeps doing its job.
No. Rootd exposes read APIs and projections. Your swap app and station software stay in place and choose what, if anything, to surface to riders and staff.
A BMS feed. Rootd's ingest is vendor-neutral — CAN, MQTT, or REST — and normalizes whatever each vendor emits into one twin, so you don't standardize hardware first.
Every query is tenant-scoped and bounded. Data is isolated per tenant by design; one operator's twin is never reachable from another tenant's context.
No. Rootd owns no hardware and operates no fleets or stations. We are software only — the cloud intelligence layer above networks other people run.
More charged packs in the right station — and more daily rider earnings retained after swap friction.
See Rootd run on your network's telemetry. We'll forecast one of your stations live.