Natural gas & power · Brazil

Questions about your natural gas processes?
Ask Polaris.

You will not regret it.

Polaris is the platform that thinks the natural gas and power market through data, intelligence and technology. You ask in plain language, and the answer comes back with a number, a source and its provenance — not from a model's memory.

24Apps
+50Intelligence models
+3kCalculation methods
< 15sResponse time
Polaris · Market Intelligence
You

What is the molecule cost at the city gate today, and what changed this week?

Before and after the trade

The desk closes the deal. The money is made at the edges.

Polaris is a BlueGas solution — a company that uses technology and innovation to think about natural gas and power from the business process outward. The focus is not the trade itself: it is everything that happens before and after it, where entropy eats time, margin and cash.

Before · The best purchase

It is not just the lowest price. It is the right purchase.

Price and grid point, yes — but also the best moment to buy, the inbound and outbound logistics, and the volume and term that make the molecule close better. The fair value in front of you before you sign, not after.

After · Fewer penalties

The imbalance before it becomes a fine.

Polaris crosses contract and scheduling data and projects the consumption trend by regression. The deviation shows up early, together with what to do about it — that is how balancing penalties and contractual fines stop being a month-end surprise.

Always · An easier operation

Less entropy between contract and allocation.

Efficiency in managing contracts, scheduling, nominations and allocations — the processes that today get solved in spreadsheets, email and rework.

Always · Leaner logistics

The best setup to receive and deliver the gas.

Beyond price, penalty and operation, Polaris builds the best logistical setup for receiving and delivering — reading the grid through hydraulics and thermodynamics, not only through the contract.

The platform

Ask. The market answers.

This is not a generic copilot bolted onto a system. It is an intelligence trained on the physics of the pipeline, the clause in the contract and the rule at the regulator.

What it is

An intelligence platform for the gas and power market, designed from the business processes outward — not an assistant glued onto a system that already exists.

How to use it

Ask in plain language, in English or Portuguese, at the depth you choose: fast, balanced or deep.

What it connects

More than 50 public sources — ANP, MME, EPE, ONS, CCEE, pipeline operators, weather and international benchmarks — plus your company's own data and whichever model you choose to run.

What it delivers

An answer with number, source and provenance — live, reference or estimated — in chat, voice, PDF or export.

Why it matters

The number always comes from the official source, with the citation attached. The model does not answer from memory — and when it does not know, it says so.

+50

intelligence models working together

Each question calls the ones that matter. These are the main domains:

Commercial Contractual Regulatory Operational Topological Hydraulic Market intelligence Pricing

+3,000 technical calculation methods sit behind them — hydraulic, thermodynamic, contractual and financial. That is what separates a plausible answer from a correct one.

Apps

Twenty-four apps (and counting…), one single context.

Each one answers a question that repeats every day. They all open inside the conversation and share the same sources and the same agent context — you never switch systems.

Market & Price

5
  • Market IntelligenceDaily sector briefing, in text and voice
  • Market IndicesBrent, WTI, Henry Hub, TTF and FX on one map
  • Molecule PricesThe NG and LNG value chain, wellhead to burner tip
  • Spread AnalysisSpreads, parities and volatility across regions
  • Liquidity & StorageWhere gas is available at each point of the grid

Operations & Grid

9
  • GMS BalancingMetering against allocation, with deviations explained
  • Smart SchedulingAssisted daily scheduling
  • Smart NominationNominate to the carrier without rework
  • Predictive BalancingThe imbalance before it happens
  • Thermal DispatchGas-fired plants and the demand they pull
  • GIS MapPipelines, plants, terminals and processing units on a living map
  • Grid SimulatorScenarios across the transport grid
  • Pipeline OptimizerRoute optimisation with file ingestion
  • NG Contract PlanningContracts and commitments on the horizon

Decision & Risk

6
  • Commercial Decision HubBuy or sell, with fair value and a recommendation
  • Hedge LabExposure, hedging and derivatives
  • Consensus PanelTrading, risk and regulatory debate — with minutes
  • Scenario SimulationPrice, supply and demand theses
  • Crisis ReplayHistorical episodes to calibrate the decision
  • Regulatory PanelANP, MME, EPE and Congress radar, with impact

Data & Reports

4
  • Dataset ExplorerThe sector's data catalogue, with filters and export
  • Unified Grid IntelligenceGas and power on the same screen
  • Report OverviewA full institutional report in PDF
  • DatalakeQuery the archive directly, without writing SQL

Just call the app by name in the middle of the conversation — it answers with a deterministic screen and the same provenance as the agent.

Natural gas is now a market of data.

Architecture

Five layers, one brain.

Robustness is not a feature: it is in skipping no step between the question and the answer.

01 The conversation One only, on desktop, on mobile and by voice — and the apps open inside it.
02 The delivery The same answer on desktop or mobile — in chat, PDF, audio, spreadsheet or straight into your system.
03 The intelligences A single brain reads the question and calls whoever knows how to answer it.
04 The sources Our own base, live sources and your documents — always anchored to their origin.
05 The data Polaris goes to where the data already lives. There is no parallel copy of your archive.

Security and governance

Every client isolated, your own key, per-person profiles and a record of what was queried. Controls aligned with SOC 2 and Brazil's LGPD.

BYOK & BYOD · your key, your data

You bring your own key and whichever model you prefer, plus your own data, which connects to the Polaris data lakes. From then on the platform answers with your contracts, your history and your house vocabulary — it stops being generic and becomes yours.

Open through A2A

Polaris talks to other agents and systems through the A2A standard. It joins the workflow you already have, instead of demanding one more screen.

Nothing invented

Price always comes from the official source, with the citation attached. The model does not answer from memory — and when it does not know, it says so.

Data & governance

The data is yours. So is the model.

The strategy is not to copy your archive. It is to go where the data already lives and hand back a decision anyone can audit afterwards.

The data stays home

Polaris connects to your archive where it already sits, and only reads. No parallel copy, no new archive to maintain.

The model runs in your environment

Your company's knowledge stays separate from every other client's, under your key. Nothing of yours feeds a shared base or a third-party model.

Every answer leaves a trail

Every figure carries where it came from, when it is from and what it passed through. Any number can be reproduced later, exactly as it came out the first time.

116Datasets in the catalogue
12Domains covered
680k+Data points
24Months of history

The sources

More than 50 public sources, plus your own.

Polaris arrives already connected to what the sector publishes. Your data goes on top of that — and it is the crossing of the two that produces the answer.

Regulation & planning

  • ANPProduction, movement, tariffs, specification, resolutions, notices and contracts
  • MME / CNPEPolicies, ordinances, guidelines and sector decisions
  • EPETen-year plans, balances, grid studies and demand scenarios
  • ANEEL / ARSEPsAssets, tariffs and federal and state power regulation

Operations & power

  • ONSLoad, generation, interchange, availability, reservoirs and constraints
  • CCEESpot price, market files, settlement and associated data
  • DESSEM / DECOMP / NEWAVEHourly, weekly and monthly horizons
  • Weather — CPTEC, ERA-5Rainfall, temperature, wind, radiation and forecasts

Gas chain & portals

  • TAG / TBG / NTSBulletins, capacities, tariffs, maintenance, flows and statements
  • Distributors and terminalsTariffs, calls, consumption, LNG and regasification
  • POC / PEGCapacities, offers, calls, reservations, contracts and positions
  • Agents and associationsProduction, studies, bulletins and aggregated sector data

Market & your data

  • Central bank FX and commoditiesFX, Brent, Henry Hub, TTF, JKM and parities
  • Data lakes and lakehousesObjects and analytical tables from your environment
  • ERP / ETRM / CRMContracts, costs, positions, customers and billing
  • SCADA and sensorsPressure, flow, temperature, metering and operating states

The first three columns are public and arrive ready. The last one is yours — and it is what turns a market overview into your decision.

Security

The key belongs to the client. So does the data.

What IT needs to hear before saying yes — in four lines on each side, and the same thing that goes into the contract.

Where your data sits

  • In BrazilProcessing and storage inside your own perimeter.
  • Under your keyYou connect your own model and set the rules of use.
  • Separate from everyoneNo client sees, trains on or influences another client's data.
  • Switchable off at any momentImmediate access cut by person, by app or by environment.

What gets recorded

  • Every queryWho asked, what was queried and what came back.
  • Where the number came fromLive, reference or estimated — always stated in the answer itself.
  • ReproducibleAny past answer can be reproduced exactly as it came out at the time.
  • SOC 2 and LGPDSecurity and privacy controls aligned with both.

Bring a question that today takes four hours.

In a 40-minute demo we connect Polaris to a dataset of yours and answer that question with a number, a source and its provenance.