MeerTech
Talk to us

MeerTech Ltd

Operating systems for the infrastructure that powers and feeds Africa.

We build the software layer for energy, climate, agriculture, health and public sector systems: offline first, auditable, and explicit about what the data does not know.

See what we have built How we build
Decision Operator action Ledger entry Indicator Unavailable Missing input MeerTech operating layer Capture Verify Compute Report Field Mini-grid Site telemetry Meter Vending events Field app Offline 00 Sensor Environment Records held on device until the link returns
Fig 01 · Field to decision Hover any stage of the operating layer to read what it does.
SectorsEnergy · Climate · Agriculture · Health · Public sector
Operating modeCapture · Verify · Compute · Report
PostureOffline first · Deny by default · Append only
BaseAbuja, Federal Capital Territory, Nigeria
01Thesis

The constraint is not intelligence. It is the field.

Models improve every quarter. What does not improve is the signal at a rural mini-grid site, the meter with no conductor specification on file, or the advisory written in a language the farmer does not read.

MeerTech builds for those conditions first, because a system that fails there was never an operating system.

High voltage substation structures silhouetted against a dusk sky

Where the software runs

Distributed assets, intermittent power, incomplete records, and a decision that cannot wait for a sync. These are the operating conditions, not the edge cases.

02Doctrine

Six constraints every system is engineered against.

They are not preferences. Each one is enforced in the database, not left to application code.

Offline first

Capture continues without a network. Sync is a design assumption, not a recovery path.

Append only

Nothing is edited in place. Corrections are new entries that reference what they correct.

No fabricated data

Where a required input is missing the system reports the absence and its reason. The database refuses the write.

Bounded AI

Model outputs run inside a deterministic envelope with a defined failure mode.

Double entry

Money moves through ledgers with idempotency keys, so a retry can never double post.

Deny by default

Access is refused unless explicitly granted, enforced at the row rather than the endpoint.

03Applied AI
Input Verified data Missing input No basis Refused Deterministic envelope Model Defined failure mode Forecast With error vs baseline Classification With confidence Inference With lineage The envelope forbids: writing to a system of record · moving money · inventing a missing input
Fig 02 · The AI envelope Schematic. No measured values are shown.

AI where it earns its place.

Solar resource and yield forecasting, crop disease classification, inference over incomplete telemetry. Each one is specific, testable, and held to a baseline.

Held to
A forecast is a capability only once it beats persistence on real data.
Published with
Its error, its confidence, and the record it was derived from.
Never
A number with no physical basis, presented as though it had one.
04Sectors

Five sectors. One engineering approach.

Pole-mounted solar panels beside a pump house in arid scrubland

Energy and grid

Mini-grid operations, metering integrity, field workforce systems and last-mile energy payments.
Lightning over an open field at night

Climate and renewables

Forecasting, storage performance and loss attribution for solar and storage assets.
Smallholder fields patterned across hills

Agriculture

Advisory in local languages, offline field capture and GIS decision support for state agencies.
Rows of medicine vials

Health systems

Offline-first capture, auditable records and supply tracking for operations under constraint.
City at night seen from above with lit street grid

Public sector data

National analytics infrastructure and field verification with lineage and immutable audit.
Multi-level road interchange at dusk with vehicle light trails

All solutions

The doctrine does not change between sectors. The failure modes do.
05Field conditions

A device offline for eleven days is not an edge case.

It syncs the way a device offline for eleven seconds does: the same reconciliation, the same audit chain, the same refusal to overwrite newer truth with older data.

Read the engineering doctrine
Device Encrypted at rest Local queue Hash chained Link down Reconciliation Idempotency keys On reconnection the queue flushes in order A retry cannot double post
Fig 03 · Capture and reconciliation Schematic.
06Evidence

Systems a client is running in production.

Grouped by deployment status, and never upgraded ahead of the facts.

01

MSIV

Deployed

Offline-first inventory management for a mini-grid operator, with scanner and label printer integration.

 02

Agricultural intelligence platform

Deployed

Crop disease detection and local language advisory for a state government client in Northern Nigeria.

 03

NMAS

Deployed

National analytics infrastructure for the National Bureau of Statistics, with methodological lineage.

 04

AditusOS

Built

The operating layer for solar mini-grid developers. Complete and tested, awaiting its first operator deployment.
Optical fibre strands glowing at their cut ends against black

Research

Eleven notes on building systems that report what they do not know.

Read the field notes

Contact

Tell us what you are trying to operate.

Talk to us About MeerTech