IntelliStream DataHub
A shared model of your operation. DataHub connects the measurements your equipment produces to the things those measurements are about, so a number in a report can always be traced back to the pump, the process and the business decision it belongs to.
Most organisations have plenty of data and very little context. A pressure reading sits in one system, the pump it came from in another, the maintenance history in a third, and the report that depends on all three is assembled by hand in a spreadsheet.
DataHub fixes the context problem. It stores your data alongside a model of what your operation actually is, the equipment, what it does, and how it all connects, so questions get answered by following relationships instead of joining five systems by hand.
Choose where to begin
The plain-language explanation, with no jargon and no assumed background.
The fourth industrial revolution explained through the three before it, and what this one actually requires.
Getting your own data out from behind vendor data models. Usually the fastest first win.
Everything in DataHub is one of four simple things. Learn them once, use them everywhere.
Build a small working model in the console, step by step.
What the four blocks add up to once live data keeps them in step. Grown, not bought.
The four ways people read these docs
What an ontology, a taxonomy and a knowledge graph actually are, what contextualization means, and how to build a model that describes your own operation.
A tour of every part of the console: data sets, the asset graph, time series, charts, events, files, millisecond data subscriptions, and relationship analysis as it lands.
Where the return actually comes from, how to pick a first project, how to measure it, what AI agents make possible, and a briefing you can take to a board.
Organisations, users, permissions, retention, backups, installation and the security posture, written for administrators, not developers.
Why this matters now
Industry 4.0 promises operations that sense, predict and adjust themselves. Every one of those capabilities depends on the same missing ingredient: a machine-readable model of your operation. An AI agent that sees a stream of numbers with no knowledge of which pump produced them, what that pump feeds, or what "normal" means for that duty cycle, is guessing. Give it a knowledge graph and it reasons faster, and with far fewer wrong answers. What AI agents make possible →
That model is the thing DataHub asks you to grow, one question at a time, and the thing it is designed to make cheap to grow. Everything else, dashboards, reports, alarms, machine learning, becomes comparatively easy once it exists.
Read the operational data problem. If you recognise your own organisation in it, the rest of this site is worth your time. If you don't, you may not need a platform like this yet.
What DataHub gives you
| One model over every silo | Historians, ERP, maintenance systems and monitoring tools all reduce to the same four building blocks. You stop learning source-system schemas. |
| Answers by traversal | "Which assets under this process saw abnormal readings during yesterday's shift, and what was raised against them?" is one question, not a week of joins. |
| Built for auditable numbers | Figures come from one queryable model and reproduce. Full lineage back to raw measurements, with data-quality flags, is on the roadmap. |
| Nothing proprietary underneath | Open source, standard components, your infrastructure or ours. Your model and your data leave in open formats whenever you want. |
- What is DataHub?: the five-minute explanation
- What is Industry 4.0?: the four revolutions, and what this one needs
- What is data liberation?: the fastest first step
- What is a digital twin?: the model, kept in step with live data
- Data subscriptions: data pushed to you the moment it arrives
- Glossary: every term on this site, defined in one line
- Frequently asked questions: the questions people ask before adopting
- Developer and SDK documentation: for teams writing code against the API