From the field
Manufacturers planning a plant-wide data foundation ask us the same things. Here are their questions, in their words, with what MaestroHub does today and a real screen from a running instance.
Where it fits
A broker moves messages; it does not know what a value means, whether it is any good, or who wrote it. MaestroHub keeps a meaning, a quality verdict and the writer on every value, permissions per topic, and a buffer when a destination is down.
No. MaestroHub reads the OPC servers you already run and connects to SCADA platforms and other brokers as sources and destinations. Nothing below it has to move on day one.
Security and IT approval
Yes. You install it on your servers, edge devices or your own cloud account, and the licence is checked offline. A new installation sends nothing to us until an administrator agrees, and air-gapped sites can switch it off.
Single sign-on over OIDC, SAML and LDAP, roles in every workspace, permissions per topic, stored credentials encrypted per installation, and TLS. MFA comes from your identity provider; there is no built-in MFA.
Running it
Collect close to the machines, standardise once in one plant structure, and distribute per use case to specialised stores.
Each site can be its own workspace in one installation. A site's connections, pipelines, topics and dashboards export as one bundle that imports into the next site, with a dry run first. Central push to many installations is not available yet.
Can we trust the numbers?
Every value is good, uncertain or bad, with a reason. A sensor that freezes is flagged even when the PLC still says good, and a source that goes quiet shows as late with the failing connection named.
Any value walks back through the pipeline run and function version to the machine address it was read from, and records who wrote it: a person, an API client or an AI agent.
Your stores and your systems
A pipeline per target writes to InfluxDB, TimescaleDB, ClickHouse and others, to PostgreSQL with pgvector, to Neo4j, and to Kafka and the lakehouse.
Every target has its own buffer on disk. An outage at one store becomes a pause, not a loss, and the buffer is delivered when the store returns.
Process flow and graph
MaestroHub publishes the plant structure, the meanings and the process flow to your graph database, keyed by the same topics, so a graph question comes back with values and quality.
Trying and buying
Yes. Download it and it runs an evaluation you can extend, with a factory simulator for live data without hardware. For a pilot we propose one line, no production writes and six weeks.
Per site, not per tag and not per user. Prices are on our offerings page.
Every answer on this page links to the docs, where you can follow the steps on your own evaluation copy.
AI agents
Agents connect over MCP with their own access token. They see only what that token allows, and every value they write records who wrote it.
Any MCP client works, including one driven by a model you host, and MaestroHub calls no model unless you configure one. Copilot Studio has a setup guide.
Questions come from manufacturers we work with and are published without names. Screens from MaestroHub on our demo plant. Full answers, with the steps: docs.maestrohub.com/questions. For volume, see the measured benchmark on the performance page.
Ask us. We answer on a running instance, and the good questions end up on this page.
