Nine engines · One architecture

THE NINE
ENGINES.

Seven Front-of-Meter engines for operating a battery in wholesale markets, and two Behind-the-Meter engines for cutting a facility's demand charges. Each is the definitive implementation of its category — purpose-built, not repurposed.

Below, each engine's real method is named plainly, as a credibility signal. We name the method — we don't publish the recipe.

Front-of-Meter · Seven engines

OPERATE THE MARKET.

01 · FoM
FENRIR

A dispatch plan built for the expected day fails on the day that isn't expected. FENRIR builds a plan that survives the bad case and still leaves room to react.

Adjustable robust optimisationDistributionally-robust (Wasserstein)Spectral risk
02 · FoM
SYNDEX

Bid as if your trades are free and you sell into a price you pushed down yourself. SYNDEX prices that impact and sizes around it.

Time-varying (Kalman) impactFat-tailed volatilityStructural-break detection
03 · FoM
KAIROS

Two things moving together doesn't tell you which caused which, or whether a third thing caused both. KAIROS answers that from your own data.

Causal discovery (CPDAG)Nonparametric independence testsLatent-confounder detection
04 · FoM
SPECTER

Treating every hour as an average hour throws away the structure. SPECTER learns the handful of regimes your grid actually lives in.

Full-covariance regime clusteringHidden-Markov dynamics
05 · FoM
CONSUL

Bidding reserves and energy with separate controllers wastes the asset. CONSUL commits them together and pre-plans the recourse for each way the day can go.

Two-stage stochastic co-optimisationCVaR recourse
06 · FoM
HELIOS

Every downstream decision needs to know how much to trust the forecast. HELIOS reports that trust as a calibrated range, not a single figure.

Covariance-aware ensembleLocally-adaptive calibrated intervals
07 · FoM
STRIX

A policy that looks great on typical days can still ruin a quarter on a rare one. STRIX concentrates its simulation on exactly those days.

Importance-sampled tail estimationStratified variance reduction
Behind-the-Meter · Two engines

CUT THE DEMAND CHARGE.

08 · BTM
PHALANX

Miss the monthly peak by a single interval and you pay for it for thirty days. PHALANX is built to not miss it, without wasting charge you didn't need.

Peak-miss tail estimationAsymmetric-cost dispatch
09 · BTM
TARSIS

Optimise one charge at a time and the pieces fight each other. TARSIS solves every part of the tariff together, to the exact optimum.

Exact mixed-integer optimisationCoincident-peak & 12-month ratchet
One architecture

THE ENGINES CHECK EACH OTHER.

The depth isn't nine separate tools — it's that they feed and cross-check one another. Five of those links, in plain terms:

SPECTER sets the context

SPECTER's read of the grid's regime tells FENRIR how hard to hedge and TARSIS which pattern to expect.

SYNDEX prices the risk

SYNDEX's market-impact estimate feeds FENRIR's worst-case dispatch, so robustness is priced to real liquidity.

HELIOS feeds the peak

HELIOS's load-and-temperature forecast feeds PHALANX, so it sees the monthly peak forming in time.

FENRIR warns of the surge

FENRIR's surge signal warns TARSIS before a demand event sets the bill.

CONSUL and TARSIS share the battery

When a site does both market and demand-charge work, CONSUL and TARSIS co-optimise the same asset as one.

ALL NINE, PLUS CHAI.
FREE FOR 90 DAYS.

No feature gating. No credit card. Proven on your own data before anything goes live.