Industrial Automation › Panda

Reads your machines, whatever they speak.

Turns machine signals into numbers people can act on: output, downtime, speed losses, quality stops. Derived values you define once, such as an availability figure or what counts as running, are calculated for you and kept alongside the raw data. It runs on a small appliance in the plant, so it keeps recording when the network does not.

In daily production across multi-site manufacturing groups

100+

Machine interfaces in the box

1

Appliance per plant

24/7

Recording, network or not

0

Internet connections required

OPC-UASiemens S7ADAMRESTPlain TCPCSVOn-premiseAir-gapped

The problem it removes

The machines already know. Nobody else does.

The data exists inside the controllers and the sensors on the line. Getting it out means a different approach for every machine generation, and what usually happens instead is a clipboard, a whiteboard and a number that arrives a day late. This closes that gap without replacing the machines.

Six ways into a machine, in the box

OPC-UA for the modern lines, Siemens S7 for the controllers, ADAM modules for the sensors on older cells, REST for the gateways that already publish, plain TCP for equipment that speaks only its own protocol, and a watched folder for lines that can only write a file.

The numbers, not the raw tags

A raw tag is not something anybody can act on. You define once what availability means on this line, or what counts as running, and the value is calculated for you and stored next to the signal it came from. Everyone downstream reads the same definition.

It keeps recording when the network does not

The recording happens in the plant, on the plant’s own hardware. A link to headquarters going down costs you a delayed report, not a hole in the shift record.

A plant is online in an afternoon

One appliance, one installation, and the collection, the storage and the dashboards come up together already wired to each other. Sites with no internet connection at all install exactly the same way.

Dashboards on the day it is installed

Standard line and shift views are there from the start, so the first conversation about the data happens in week one rather than after a reporting project.

Old lines are not left out

The equipment that cannot be modernised this year is usually the equipment losing the most time. A watched folder or a plain socket is often enough to bring it into the same picture as the new line next to it.

What is in the standard set. The interfaces above (OPC-UA, Siemens S7, ADAM, REST, plain TCP and watched CSV folders) are the ones that ship ready to configure. MQTT, Modbus and serial connections are not in the standard set today: where a line needs one, it is built as part of the project. Ask early and we will tell you which of your machines fall into which group.

What it does

One way in for every generation of machine.

OPC-UA

The modern machines, the standard way

Subscribes to the values that matter and records changes as they happen, which is what most equipment built in the last decade expects.

Siemens S7

Controllers read directly

Reads controller memory directly, without an extra gateway to install in the cell and keep alive.

ADAM modules

The sensors bolted onto older lines

Standard signal and digital inputs from the measurement modules that are already installed on plenty of older cells.

REST

Whatever already publishes

Where a cell gateway or a machine already offers an interface over the network, it is read on a schedule you set.

Plain TCP

Equipment that speaks only its own language

For machines with a proprietary protocol and no driver, the raw connection is read and interpreted according to the description you provide.

Watched folders

Lines that can only write a file

A folder is watched and new rows are read as they land, which is often the shortest route to including a legacy line in the same picture.

End to end

From a signal on the line to a number you can use.

Machines, sensors and existing systems are read over the protocols they already speak: OPC-UA, Siemens S7, Modbus TCP, MQTT, REST and files. Panda runs a collector for each, calculates the values you define yourself, and works out what state each machine is in. Out of it comes a history you can query, a live picture on the screens, and a stream other systems subscribe to.

How it runs

One appliance per plant, installed in a single pass, carrying the collection, the storage and the dashboards. It needs nothing from the internet and nothing from headquarters. Sites with their own cluster can run the same software there. Each plant keeps its own certificates and its own address, and rolling out to the next plant is the same procedure again.

Appliance per plantPrivate KubernetesAir-gapped installLocal certificatesDashboards includedMulti-site rollout

Works with the rest of the range

The collection sits between the machine control that produces the events and the order system that turns them into completions and traceability. The same recorded values feed a group reporting cockpit, and an automation can act on a deviation while the shift is still running.

How to start

One line, one week, real numbers.

We connect a single line and let it record for a week. What comes back is normally enough to settle the argument about where the plant is actually losing time.