Cloud or on-prem · Guided demo

Cloud SCADA,
AI-native.

Monitor and control your industrial facilities from the cloud or on-prem. Build dashboards and synoptics by talking to the AI or with the visual editor — you choose how to work.

  • No-strings demo
  • Multi-tenant
  • Cloud or on-prem
lymon.lybits.es/designs/north-plant

Consumption · North Plant

Last 24 h · live

Online

Consumption

128 kW

+4.2%

Daily cost

412 €

−1.8%

Alarms

2

active

Active power (kW)

Power factor

0.97

"Add a panel with consumption per line over the last 8 hours"

Generating panel…

Two paths, one platform

Talk to the AI or click. Without giving up anything.

Everything you can ask for by chat you can also do with the visual editor. The plant operator does not need to chat; the engineer who prefers AI goes ten times faster. The platform serves both.

AI chat (MCP)

Connect Claude or another model over MCP and ask for what you need in natural language.

  • Create and edit designs by describing them
  • Query history and evaluate formulas
  • Claude by default or your own key (BYO)

Interactive UI

A complete web portal with a drag-and-drop editor. Without writing a single line.

  • Visual editor for dashboards and synoptics
  • Variable catalogue with search and filters
  • Manage agents, alarms, users and plans

Capabilities

Everything a connected plant needs.

From the data in the PLC to the panel on screen and the command back to the equipment. One platform, no custom integrations.

Edge agent

A Rust binary under 10 MB that connects Modbus, OPC-UA and MQTT, with a durable buffer and recovery from outages.

Designs and widgets

A unified editor for dashboards and synoptics, a formula engine and a sandboxed Widget SDK for your own components.

ISA-18.2 alarms

Live alarm management, catalogue and performance metrics per the standard. No ungoverned event lists.

Traceable control

Write-back to equipment with an allow-list, signing and OPC-UA. Every action is logged.

Multi-tenant by default

Row-level isolation (RLS), audit logging and access with Microsoft Entra. Multi-tenant, not an add-on.

Your choice of LLM

Claude by default or your own key: OpenAI, Gemini, Mistral or local Ollama. With routing and a fallback chain.

How it works

From the shop floor to the screen in three steps.

1

Install and connect

Deploy the agent next to your equipment and point it at your PLCs and meters. Modbus, OPC-UA or MQTT.

2

Continuous ingest

The agent uploads readings to Lymon exactly-once. If the network drops, it buffers and retries without losing a single sample.

3

Visualize and control

Build designs by chat or click, define ISA-18.2 alarms and send commands back to the equipment with full traceability.

Plant

PLCs, meters and sensors speaking their usual protocols.

ModbusOPC-UAMQTT

Lymon agent

A lightweight edge binary: samples, buffers and retries through network outages.

Rust · <10 MBDurable bufferExactly-once

Lymon

Ingest, time series, formula engine and MCP gateway. Cloud or on-prem.

Multi-tenantTimescaleEdge / air-gapped

People and agents

Dashboards and synoptics, AI chat, alarms and traceable control.

DashboardsChat / MCPAlarms

Cloud or on-prem

The same platform, wherever you decide.

Lymon runs the same in the cloud as inside your network. The same images, no branches or surprises. Built for the edge from day one, even with no internet access.

  • Managed cloud

    Multi-tenant on Azure, with nothing for you to maintain.

  • Edge / on-prem

    K3s, Docker Compose or appliance. Your data never leaves home.

  • Air-gapped

    Offline installation via a self-contained bundle.

< 10 MB

Edge agent in Rust

1k/s

Samples per second, sustained

Same

Images on cloud and edge

RLS

Per-tenant isolation

Plans

Start free. Grow when you need to.

Ask us for a demo and we set up a trial workspace for you. No card required.

Trial

Recommended

Free for 14 days

To try the platform end to end.

  • A workspace with sample data
  • Edge agent and connectors
  • AI designs and visual editor
  • Chat with Claude included

The trial lasts 14 days. Without an upgrade the workspace is locked and, after 15 more days, permanently deleted along with all its data.

Request a demo

Production

Custom

To run your plant in the cloud.

  • Higher limits for series and agents
  • ISA-18.2 alarms and control
  • Your own LLM key (BYO)
  • Roles, audit and support
Talk to sales

Edge / Enterprise

Custom

On-prem, air-gapped or hybrid.

  • Deploy in your network (K3s / appliance)
  • Offline installation
  • SSO and dedicated isolation
  • Custom support and SLA
Talk to sales

FAQ

What people usually ask.

Do I need to know about AI to use Lymon?
No. Everything the platform offers can be done with the visual editor and the portal, by clicking. AI is one more path to go faster, not a requirement.
Can I use it without the cloud, inside my network?
Yes. Lymon is built for the edge from day one: K3s, Docker Compose or appliance, even installations with no internet via a self-contained bundle. They are the same images as in the cloud.
What equipment and protocols does the agent connect?
The edge agent speaks Modbus, OPC-UA and MQTT with your PLCs, meters and sensors. It buffers locally and retries through network outages without losing data.
Who can see my data?
Each tenant is isolated at the database level (RLS) and every sensitive action is written to the audit log. If you prefer, deploy on-prem and your data never leaves your site.
Which AI models can I use?
Claude comes by default. You can bring your own key (BYO) from OpenAI, Gemini or Mistral, or run a local model with Ollama. The router manages providers and the fallback chain.
What makes Lymon different?
Lymon combines the best of a mature industrial platform — a formula engine, Widget SDK and a unified Query API — with multi-tenant cloud, edge, AI via MCP and a complete portal. AI-native from day one.

Shall we see it with your data?

We walk you through the platform with a case like yours and set up your trial workspace. No strings attached.