Solantir turns raw, confusing, and overwhelming data into something a person can actually understand and act on — for students trying to see where their grades are heading, and for organisations that need their data to think alongside them.
/soʊˈlæn.tɪər/
Solantir
A fusion of Solon — the Athenian statesman remembered for restoring order and wise judgment out of civic chaos — and tir, a root meaning "to see" or "to perceive."
To see clearly, and through seeing clearly, bring order to what was previously confusing.
Data is rarely the problem. What people actually struggle with is data that hasn't been made legible — numbers without narrative, records without relationships, signals buried inside noise. Solantir exists to close that gap.
We are a South African data infrastructure and applied AI company built around one core capability: taking raw, confusing, or vague data and turning it into something a person can understand and act on — whether that person is a student checking their grades or an executive reading a supply chain risk report.
Raw data arrives messy, incomplete, and scattered across systems that were never designed to talk to each other. Solantir's engines ingest it exactly as it is — no assumptions about what it should look like.
Three infrastructure engines — Omnis, LumeX, and Spector — clean, structure, and interrogate that data: finding patterns, detecting anomalies, and mapping relationships that weren't visible before.
The result surfaces through Lume and Kestrel as ranked, specific, decision-grade intelligence — not another dashboard to interpret, but a clear answer to "what should I actually do next."
Underneath Lume and Kestrel sits Solantir's real technical asset — a set of infrastructure engines, each responsible for a distinct part of turning raw data into usable intelligence.
Pattern Recognition & Prediction
Domain-agnostic by design — Omnis isn't built to understand one kind of data. It finds patterns, detects anomalies, and generates forward-looking insight across whatever structured data it's given, regardless of industry.
Predictive Model for Academic Data
The analytical core beneath Lume. LumeX takes a student's academic data and generates forward-looking insight about where their performance is heading — the model underneath every forecast Lume shows.
Data Construction & Orchestration
The layer responsible for structuring and organising raw data before it can be meaningfully analysed. Solantir's newest engine — still in research — built to eventually power Sable.
Three engines, three live demonstrations, zero mockups. Hit activate on each one and watch raw, messy, real-world-shaped data resolve into something you can actually use.
01 Omnis
Below is a small supply network: one factory, three warehouses, three stores. Somewhere in it, a shipment pattern has quietly gone off-script. Activate Omnis and watch it find what a spreadsheet would miss.
Technical readout
Anomaly detected: WH-GQB → STORE-07. Shipment volume +286% vs. 90-day rolling baseline. Pattern confidence: 92%.
In plain language
The Gqeberha warehouse is suddenly sending almost 3x its normal shipment volume to Store 07 in Mitchells Plain — a route that barely existed a month ago. Worth checking before stock runs short somewhere else in the network.
02 LumeX
Three terms of real marks, one student. Activate LumeX and watch it forecast Term 4 for every subject — then tell you exactly where to spend your next study session, and why.
| Subject | Term 1 | Term 2 | Term 3 | Term 4 (Predicted) |
|---|---|---|---|---|
| Mathematics | 68% | 61% | 54% | — |
| Physical Sciences | 72% | 74% | 70% | — |
| English Home Language | 80% | 78% | 82% | — |
| Life Sciences | 65% | 68% | 71% | — |
| Accounting | 74% | 73% | 75% | — |
Technical readout
LumeX projection: Mathematics trending −14pts over 3 terms (68→61→54). Term 4 forecast: 47%. Confidence: 91%. Risk classification: High.
In plain language
Your Maths mark has dropped every term this year. If nothing changes, LumeX predicts you'll land in the high-40s next term. Everything else is stable or improving — Maths is where your next study session should go.
03 Spector
Every supplier names their fields differently, formats dates differently, and makes different typos. Activate Spector and watch four messy records resolve into one clean, structured dataset.
Record 01
Record 02
Record 03
Record 04
Technical readout
Spector normalised 4 records: resolved 4 field-name variants to one schema (supplier_name, quantity, unit_price, delivery_date), corrected 4 invalid or ambiguous dates, repaired 3 malformed currency values, fixed 1 non-numeric quantity.
In plain language
Four messy supplier records — different field names, broken dates, a typo instead of a number — became one clean, structured dataset in seconds.
Lume
Built for South African matric, university, and NSFAS-funded students. Lume takes your actual marks and tells you three things: where your grades are heading, your single biggest academic risk right now, and exactly what to focus on to address it — backed by a personalised, calendar-scheduled study plan.
Kestrel
Kestrel runs on Omnis, passing your operational data through a six-layer pipeline — clean, normalise, derive, detect, map, rank — to surface prioritised intelligence. Specialised for supply chain risk, flexible enough for any operation that can't afford errors in its data.
Lume is built for high school students preparing for matric, university and college students, and specifically for NSFAS-funded students, for whom maintaining academic performance is a requirement for keeping their funding.
Kestrel is built for institutions where errors in data carry real consequences: procurement officers who need visibility into supplier risk, analysts detecting patterns across large and complex datasets, and executives who need decision support grounded in their organisation's actual operational reality — not a generic dashboard.
Different audiences, same underlying premise: turn what's vague into something you can trust and act on.
Built, not bought
Lume
Academic forecasting · Personalised study plans · NSFAS-aware
Kestrel
Supply chain intelligence · Six-layer pipeline · The Workshop
Sable — in research
Powered by Spector · Order from high-entropy data · Coming
Whether you're evaluating Kestrel for your supply chain, want Lume for your institution's students, or are facing a data problem the market hasn't solved yet — we want to hear about it.
Kestrel · Custom infrastructure · Supply chain intelligence
Schools · Universities · NSFAS-aligned programmes