A mountain valley beneath an open sky
RAIN / Site intelligence SIMULATED SITE

NORTH RIDGE / SITE OVERVIEW

Every signal. In view.

↳ Local runtime
Power output1.84 MWIllustrative reading
Signals observed128Processed on-site
Active advisory01Inverter temperature
Generation / last 12 hoursRecorded
06:0012:0018:00
Advisory recordedRead-only · Signed audit↗

AI that learns on-site.
Data that never leaves it.

DOM is edge AI hardware and software that runs inference and retrains models directly at your facility, so your operational data never has to leave the premises.

Explore the platform Discover the system

Supported by the Claude Network Program

SCROLL TO FOLLOW THE SIGNAL
01 / THE PROBLEMSkip to how it works ↘

More data.
Still missing the picture.

ILLUSTRATIVE SCENARIOS
EQUIPMENT SIGNAL

A moment that matters.

A brief deviation
EquipmentSite average

More telemetry. Less of the detail.

SITE ACOASTAL

Same equipment.

Salt air · variable wind
SITE BINLAND

Different reality.

Heat · seasonal dust

A shared model isn’t a shared reality.

Your facilityStill generating signals
×
Cloud modelConnection unavailable
Insight waiting on the network

Your operation can’t wait for a connection.

Scroll to explore

Intelligence needs to live closer to the operation.

02 / HOW IT WORKS

From plant signals.
To local insight.

01 / CONNECT

Read the signals
you already have.

Connect to existing equipment through read-only access. Raw operational data stays at your facility.

Your signals, captured locally.
02 / LEARN

Learn the way
your site runs.

Run inference and retrain locally. Evaluate each candidate against the current model before it can take over.

A learning loop grounded in your site.
03 / REVIEW

Give operators
the evidence.

Review advisories with their model version and recorded history. Your team decides what happens next.

Insight for people. Control stays with them.
LOCAL DATA · READ-ONLY ACCESS
Explore the hardware + software
03 / THE SYSTEMAT YOUR FACILITY

Hardware and software.
One local system.

The RAIN Node hosts the runtime and your view of the site. Your operators decide. Your control systems remain separate.

ON-SITE INFRASTRUCTURE

The local compute layer

Compute, storage and connectivity at your facility.

Conceptual architecture
RUNS ON THE NODELOCAL

RAIN Runtime

Local models. Evaluated updates. Recorded history.

  • Inference & learningON-SITE
  • Model evaluationGATED
  • Signed historyRECORDED

Local console

ILLUSTRATIVE

Review advisories, model versions and system health.

SITE SIGNAL
Advisory for reviewModel version + recorded history
READ-ONLY VIEWOperator decides ↗

Working against a simulated plant. Preparing for a first field pilot.

Connected intelligence

Your systems.
One intelligence.

Bring your existing infrastructure into one local intelligence layer. From the first signal to the next insight, everything stays on-site.

Inside your facility

Signals converge. Control stays yours.

One continuous, local loop. Conceptual view.

From connection to clarity

Bring PCS, BMS, EMS, and HVAC signals together through read-only adapters. The connection observes your site while keeping control systems separate.

Existing infrastructure

Bring telemetry into a local record of your facility. Models learn from operating patterns in the simulated plant, building toward site-specific learning.

Site-specific learning

Run inference and evaluate model updates on the RAIN Node. Keep the learning loop close to your equipment and your raw signals on-site.

Local inference

Review advisories, model versions, and signed history in the local console. Your team makes the decisions; the system provides the context.

Operator insight

A local intelligence loop, designed around your site.

Read-only by design
04 / USE CASESSTARTING WITH RENEWABLE ENERGY

Start with a real
operational question.

For asset owners, operators and maintenance teams. Start with one asset and agree what useful insight would look like.

Explore all use cases

SOLAR / PROPOSED APPLICATION

Which inverter is behaving differently?

Learn from inverter and tracker signals to surface changing behavior that can disappear inside aggregated performance reports.

Inverter telemetryTracker behavior
Explore this use case

05 / BUILT FOR OPERATIONAL TRUST

Local intelligence.
Clear operational boundaries.

01

Raw signals remain local.

The inference and learning loop uses local data. Raw telemetry is kept on-site; production uplink transport is not yet implemented.

02

Advice, with no actuation.

The system observes and advises. Operators retain authority over equipment, control systems and any action taken.

03

Decisions have a history.

A signed, linked history records inferences and model changes, including rejected updates. It is evidence to inspect, not a claim of certification.

06 / PROGRESS & OPPORTUNITYRAIN · PRE-SEED

An ambitious system.
A focused first step.

We’re building on-site intelligence for facilities where raw operational data needs to stay local.

Renewable energy is our starting point. A battery-storage simulator is the current foundation; a scoped field pilot is the next validation milestone.

  1. TODAY / SIMULATOR

    A working local loop.

    Telemetry capture, local inference, retraining, model evaluation and audit history implemented against a simulated battery plant.

  2. NEXT / FIELD VALIDATION

    Prove it at one site.

    Validate hardware, signal quality, alert usefulness and operating stability with a design partner. No field deployment is claimed.

  3. THEN / REPEATABILITY

    Make the next site easier.

    Turn pilot learning into a repeatable installation and support process. Extend to additional asset types as adapters are validated.

What we need to prove

Useful insights for operators. Reliable operation on-site. A repeatable deployment process. Commercial value established with customers.

Read the investor overview

07 / QUESTIONS, ANSWERED

Before we
step on-site.

Does RAIN control my equipment?+

No. RAIN is observe-only. It reads telemetry and produces advisories. There is no actuation path back into your plant control or safety systems.

Does the node need a cloud connection?+

The local inference and learning loop is designed to operate without runtime network egress. Raw plant telemetry stays on-site. The current simulator uplink is a local no-op; a production transport is not implemented yet.

Can we deploy RAIN today?+

The system works against a simulated plant and has not been field-deployed. The next step is a scoped pilot, with equipment access, signal quality and success criteria agreed in advance.

What does a first pilot involve?+

Define the asset and question, inventory available signals, agree a read-only boundary, then validate results against site conditions. Hardware configuration and commercial terms depend on that scope.

What happens when a new model is worse?+

A candidate must beat the incumbent on held-out data before promotion. Otherwise the incumbent stays. The rejected candidate's evaluation is still recorded in the audit history.

08 / BUILD WITH RAIN

Bring a real problem.
Help shape what comes next.

For operators exploring a pilot and investors evaluating the next stage.

FOR OPERATORS & DESIGN PARTNERSBuild your pilot brief

Define the asset, the question and what success should look like.

FOR INVESTORSExplore the opportunity

Understand the product thesis, current stage and next milestones.

Briefs download locally. No information is submitted.

START WITH THE RIGHT QUESTION

Your pilot brief.

Describe the site and the signal you want to understand. Download your brief to share with your team.

Nothing is submitted or sent. This creates a text file on your device.