Your factory, in a language AI understands.
MaestroHub is the industrial context platform: it connects every machine and business system in your plant and turns their data into one governed language, so your people and AI can see the plant, understand it and act on it, inside your rules and on your own infrastructure.
Three products, one platform
MaestroHub has three products. These docs are mainly about Context Engine, the one you can run today. This is the documentation for release 2.6.5; the 3.0 documentation is published ahead of its release.
- Context Engine Live2.6.5
See the plant. Connects PLCs, SCADA, historians, databases, files and clouds, gives every value one address, a schema, a quality and where it came from, moves it with pipelines that keep data when a link drops, and serves it over MQTT, REST, dashboards, alerts and an MCP server for agents. Sold in two editions, Foundation and Enterprise. - Knowledge Graph Coming soonnot part of 3.0
Understand it. The plant as things and the links between them: what feeds what, what controls what, what stops if this stops, and the plant as it was on any past day. - App Studio Coming soonnot part of 3.0
Act on it. A domain expert describes the application they need, and it runs on the whole platform inside that person's permissions.
AI agents work with all three through one governed door. Bring your own MCP client (Claude, Cursor, VS Code and others) today; the Maestro agent inside the product is coming soon. Either way an agent sees only what the person it acts for may see, and every step lands in the audit trail.
The whole platform, on one page
Follow the numbers, from the plant to the people, agents and systems that use its data.
1Your plantAny vendor, any age
- PLCs and controllers, CNC machines and robots
- SCADA, historians, MES and ERP such as SAP
- Databases, files, cloud and IoT services
2Context EngineSee the plantLive
- Connect: PLCs, SCADA, databases, files and clouds (90+ connectors in 3.0)
- Give it context: one namespace, schemas and units, quality and provenance on every value
- Move it: pipelines and functions, store and forward when a link drops
- Serve it: MQTT and REST, dashboards and alerts, an MCP server for agents
3Knowledge GraphUnderstand itComing soonnot part of 3.0
- Every thing in the plant and how it relates
- Ask in plain words, get the path back
- The plant as it was on any past day
4Agents and App StudioAct on itLiveagent accessComing soonApp Studio
- Your own MCP client, or the Maestro agent (coming soon)
- Apps a domain expert describes, App Studio builds
- Every action inside one person's permissions
5Governed write-backBack to the plantLive
- Actions return to the plant as a named person, checked first and recorded
- A data path, not a control loop
I want to
| I want to | Start here |
|---|---|
| Try MaestroHub in minutes, with no hardware | Get started with the Digital Factory Simulator |
| Connect a machine or a business system | Connect |
| Give my plant one namespace with schemas and history | Unified Namespace |
| Move data to a database or a cloud, and keep it when the link drops | Orchestrate and store and forward |
| Build a dashboard or an alert | Dashboards and alerts |
| Let an AI agent use my data safely | MCP integration |
| Decide who can do what | Identity and access |
What MaestroHub is not
Said first, so nothing surprises you later. MaestroHub is not a SCADA or HMI, not a control system, not an MES and not a hosted cloud service: it runs on your own edge box, virtual machine or Kubernetes cluster, above the OT tools you already own. It is not a long-term historian either. History is an operational window that you set, the same in every edition: 24 hours by default on the embedded store, and 1, 3 or 7 days, or keep forever, on TimescaleDB. Long-term history goes to your system of record through connectors. Read the full picture.