How we decidethe whole method is on the Solution design page

The radio is a decision, not a default. Five questions decide it.

  1. 01Where is the asset?
  2. 02Does it move?
  3. 03How often must it speak?
  4. 04What powers it?
  5. 05Who should own the network?
The decision table and six worked examples →
Two things we do

We advise.

Four fixed engagements, from a two-week assessment of what your assets already report to the design of a pilot with its acceptance checks written in advance. Each ends in a document you keep, whether or not anything is installed.Four fixed engagements, from a two-week assessment of what your assets already report to the design of a pilot with its acceptance checks written in advance. Each ends in a document you keep.

We build HOLOIN.

One installed platform that holds vehicles, farms, buildings, remote sites, solar plants and charging depots as one live record, learns each asset's normal with statistics a person can read, and turns a change into a work order for the person responsible. Licensed per installation, counted in devices.

— Where an integrator serves your region, they install, support and answer the phone. We never compete with them for your services, and we never write custom software.— Where an integrator serves your region, they install, support and answer the phone. We never compete with them, and we never write custom software.

The four engagementssource: engagement terms

Fixed scope. A document at the end. Nothing touched.

Each is small enough to decide on quickly. Most lead to a HOLOIN pilot; some end with the honest answer that you are not ready yet. Where an integrator serves your region, their engineer is on site with us. Each row opens to its full terms.Each is small enough to decide on quickly. Most lead to a HOLOIN pilot; some end with the honest answer that you are not ready yet.

01

Asset connectivity assessment

2 weeks · 1 day on site · start here

What you run, what it already reports, over which protocols and networks, where that data goes now and who sees it, and what can be read today without touching control.What you run, what it already reports, over which protocols and networks, where that data goes now and who sees it, and what can be read today without touching control. One site or one fleet.

Full terms →

One day on site with your OT and IT people; the rest remote. Your vendor security questions come into it, so the blueprint that follows already knows what your IT will ask. One site or one fleet.

You keep
A connectivity map, asset by asset: readable today, readable after a configuration change, not readable without new hardware; the gaps that matter; one page on what a read-only installation shows on day one.a connectivity map, asset by asset; the gaps that matter; one page on what a read-only installation shows on day one.

02

OT/IT blueprint

3 weeks · 2 days on site

How field data reaches your operations without putting control at risk: read-only first, which networks are segmented from which, where the data lives and stays, who sees what.How field data reaches your operations without putting control at risk: read-only first, segmentation, where the data lives and stays, who sees what, upgrades and backups, what your security review will want written down.

Full terms →

How upgrades and backups are done, and what your security review will want to see written down.

You keep
The blueprint your IT and OT teams both sign: connection paths, trust boundaries, the data-residency answer, and the operating rules for commands, if you ever want any.the blueprint your IT and OT teams both sign.

03

Your own operating model

2 workshops · 2 days on site

The workshop that writes down your operating model: your assets and the places they belong to, the measurements that matter for each, what “normal” means for it, which changes deserve a person, and who that person is.The workshop that writes down your operating model: your assets and the places they belong to, the measurements that matter, what “normal” means for each, which changes deserve a person, and who.

Full terms →

A record of your operation that you own and can read.

You keep
The model written out: homes, roles, measurements, what “normal” means for each, which changes deserve a person, and who. It is the configuration of a HOLOIN installation, and it is yours either way.the model written out. It is the configuration of a HOLOIN installation, and it is yours either way.

04

Pilot design

1 week · remote

One site, a scope, the people, and the checks that decide it in advance: what counts as proof, measured how, by when.

Full terms →

A pilot designed this way ends with a decision, not an opinion.

You keep
The pilot plan with its acceptance checks, the devices and hardware it needs, who does what, and the price of the installation that follows if it passes.the pilot plan with its acceptance checks, the hardware it needs, and the price of the installation that follows if it passes.

— not offered, by choice: custom software · staff on hire · anything called "AI development"

How an asset becomes livevalues: demo run 6b741936b74193

Five steps, and the asset is live. None of them touches control.

HOW AN ASSET BECOMES LIVE VALUES: DEMO RUN 6b74193 OT · WHAT YOU RUN YOUR SERVER · ONE RECORD IT · YOUR SYSTEMS AND PEOPLE Truck 7tracker · cellular · Codec 8 Feeder meter BModbus TCP · FC 03 AHU 2 controllerBACnet/IP · COV Room 104 sensorLoRaWAN · fPort 1 Charger 3OCPP 1.6 · WebSocket ERPwebhooks · API keys by scope IdentityOpenID Connect · roles Mail serveryours · notifications Technician on dutyphone · works offline reads · never writes Truck 7 · speed Feeder B · kW AHU 2 · supply °C Room 104 · CO₂ Charger 3 · kW 42 km/h 238 kW 13.8 °C 640 ppm 7.2 kW 1 520 ppm above its usual since 14:02 · usual 640 ± 90 WORK ORDER #318 → phone trend at this rate: reaches the feeder's limit on 14 Nov feeder B powers the controller and the charger · 12 alarms held behind 1 cause PROPOSED RULE · FROM THIS PLANT'S OWN FAILURES AHU 2 current ≥ 12 A · caught 5 of 6 past failures · 3× chance · a person confirms mirror knows its normal sees its trends knows its dependencies recommends What you already run speaks its own protocols. Your systems and people sit on the other side. Nothing between them yet. HOLOIN reads across the line and never writes. The connection is proven read-only before anything else is built. Every asset becomes one row of one record, on your server, with its live reading and its history. It learns each asset's normal for this hour of the week: a median and a spread you can read. No model guessing. A reading leaves its band. The record names it, and a work order reaches the technician's phone. A fault found before the shift, not after. Trends run forward at their own rate, to a date. A part is ordered before the limit, not after the failure. It learns which asset depends on which. One upstream cause holds back a dozen alarms. One visit, not twelve. From its own history it proposes rules with their backtest, and a person confirms. Chirp advises the design. HOLOIN runs it on your servers. 01 · TWO WORLDS 02 · THE TRUST LINE 03 · ONE RECORD 04 · KNOWS ITS NORMAL 05 · NOTICES A CHANGE 06 · SEES ITS TRENDS 07 · KNOWS ITS DEPENDENCIES 08 · RECOMMENDS
Eight beats, 32 seconds, then it holds its last frame. Every value is from a recorded demo run; nothing is projected. Still when your browser asks for no motion.
01

Read

reads · never writes

The devices already on your assets are read over the protocols they speak: cellular trackers, LoRaWAN, Modbus, BACnet, OPC UA, OCPP, SNMP. Across a line that reads and never writes; the read-only state is proven by a test, not promised.The devices already on your assets are read over the protocols they speak, across a line that reads and never writes.

No new hardware to start. Your OT team keeps control; your IT team gets a boundary it can sign.No new hardware to start; your OT team keeps control.

02

Record

one row per asset

Every asset becomes one row of one record on your server, with its live reading, its history, its place and its people. One record across depots, sites and hauliers, not one dashboard per vendor.Every asset becomes one row of one record on your server: live reading, history, place, people.

One place to answer “what is wrong now” for everything you run.One place to answer “what is wrong now”.

03

Normal

usual 640 ± 90 ppm

The record learns each asset’s usual for this hour of the week — a median and a spread over 28 days — written out so a person can check it.The record learns each asset’s usual for this hour of the week, written out so a person can check it.

Alarms that mean something: “too high” is measured against this asset, not a catalogue number.Alarms measured against this asset, not a catalogue number.

04

Change

above its usual since 14:02

A reading leaves its band, a trend runs toward a limit at its own rate, an upstream fault explains a dozen silent devices at once. The record says which, in a sentence with its evidence.A reading leaves its band, a trend runs toward a limit, an upstream fault explains a dozen silent devices at once. The record says which, with its evidence.

A fault found before the shift, a part ordered before the failure, one visit instead of twelve.A fault found before the shift; one visit instead of twelve.

05

Person

work order #318 → phone

A work order reaches the technician’s phone, offline or not; the result reaches your ERP through the interfaces you name. Nothing acts on a machine unless you switch that on, with a written reason, and it counts as done only on the machine’s own evidence.A work order reaches the technician’s phone, offline or not; the result reaches your ERP. Nothing acts on a machine unless you switch that on, with a written reason.

People and systems you already have, better informed. No new silo.People and systems you already have, better informed.

— each step shown on a live installation at holoin.ai/how-we-verify

HOLOINrelease 1.5.2 · source: holoin.airelease 1.5.2

One record of everything you run.

Installed on your server or on a dedicated server we run for you. On a server we run for you, our access to your data is off by default: your administrator grants it, for a set time, and every action is logged against both identities.

01

Assets

six verticals live

vehicles · farms · buildings · remote sites · energy plants · EV depots

the six verticals holoin.ai lists as live

02

Reads

a line that reads, never writes

cellular trackers · LoRaWAN · LwM2M (NB-IoT, LTE-M) · Modbus TCP · SunSpec · BACnet/IP · OPC UA · OCPP 1.6 / 2.0.1 · SNMP · MQTT · Sparkplug B · CoAP · HTTP

the protocols holoin.ai lists as live

03

Learns

median · spread · trend · backtest

each asset’s usual by hour of week (a median and a spread), its trends at their own rate, its dependencies, and rules proposed from its own history with their backtest; a person confirms every one.

Shown on a live installation at holoin.ai.

04

Writes

off by default

nothing, unless it is switched on for that site, with a written reason; a command counts as done only on the device’s own evidence.

The factory profile ships a read-only attestation with every release.

05

Licence

counted in devices, never people

annual, per installation: the platform, the verticals you run, and a device tier.

No per-seat fees; no charge for the people who look.

Recordevery line with its sourcewith sources

A company that has built these solutions.

Demonstrated

2021

A LoRaWAN smart-city system demonstrated to the Punjab Information Technology Board. Under a memorandum of understanding signed in April 2021, Chirp built and demonstrated a working LoRaWAN system for the province’s smart-city programme in Lahore.Under an MoU signed in April 2021, Chirp built and demonstrated a working LoRaWAN system for the province’s smart-city programme in Lahore.

— The Nation · 2 Apr 2021
Named under the parties’ own public announcement.

Delivered

2020

A field tool for commissioning LoRaWAN sensors. For an IoT product company: an Android app that connects to a sensor over Bluetooth, walks the installer through deploying it on a LoRaWAN network and registers it with the cloud.For an IoT product company: an Android app that walks the installer through deploying a sensor on a LoRaWAN network and registers it with the cloud.

— the client’s review · Clutch
The client is unnamed; references on request.

Architected

2018

An end-to-end LoRaWAN water-metering system for a town’s water utility. Before Chirp, its founder was the system’s IoT architect: meters in the pits, the gateways, the network server and its device provisioning, the sensor decoders, the head-end, the utility’s own dashboards.Before Chirp, its founder was the system’s IoT architect, from the meters in the pits to the utility’s own dashboards.

Finalist

2019

ChirpyWAN, Chirp’s own deployment tool for LoRaWAN networks, a finalist in Momenta’s LoRaWAN Startup Challenge. One of five finalists, chosen to pitch in Berlin on 13 June 2019. The application built, tested and validated a LoRaWAN network in the field — a network scan, a built-in sniffer, power-level, data-rate and channel validation, range tests — with coverage surveys, installation and end-to-end testing as the services around it. Each finalist received a year’s Adopter membership of the LoRa Alliance (the Alliance’s letter of 3 June 2019, on file). No source names the winner, so this line stops at finalist.One of five finalists, chosen to pitch in Berlin on 13 June 2019. The application built, tested and validated a LoRaWAN network in the field: a network scan, a built-in sniffer, power-level, data-rate and channel validation, range tests. Each finalist received a year’s Adopter membership of the LoRa Alliance (the Alliance’s letter of 3 June 2019, on file). No source names the winner, so this line stops at finalist.

— Momenta · 7 Jun 2019
Named in the organiser’s own public announcement.

Reviewed

2021–23

4.9 of 5 from four verified client reviews of our engineering work.

— Clutch · 4 reviews
Verified by the review platform, not by us.

— projects are named only with the customer's written permission or the customer's own public announcement; references on request

Contact

Forty minutes, free. You leave knowing whether an assessment is worth doing.

The five questions answered for one site, and our view in writing. If you ask, you see the platform running on a demo installation; if you don’t, you won’t. Call or write; you reach the founder, not a sales desk.

a few sentences on what you run and what you want to know