Connect
Point TagLake at an OPC UA server, an MQTT broker, a SQL Server database or a REST endpoint and it reads from it directly. There is no separate connector to install, and no broker added just to translate.
Industrial data platform
Many signals.One structure.Operational intelligence.
TagLake reads your plant directly over OPC UA, MQTT and Sparkplug B, SQL Server and REST, and gives every value one place in a Unified Namespace (UNS) your engineers design. TagLake IQ sits on top, answering questions in plain language and adding descriptive and predictive analytics.
TagLake Core is available now. TagLake IQ is in development.
TagLake Core reads these directly
- OPC UA
- MQTT
- SQL Server
- REST
- 01 Plant systems Where the signals come from DCSPLCSCADAHistoriansDatabasesMES
- 02 TagLake Core One structure Available now
- 03 TagLake IQ Operational intelligence In development
Connect. Structure. Intelligence.
In that order. Each step is useful on its own.
Structure
Every value becomes one record, with its unit, both timestamps, quality and source trail. Your engineers then contextualise it in a Unified Namespace (UNS) they design, so the plant has one shape whichever system a value came from.
Intelligence
TagLake IQ sits on Core. It answers questions about the plant in plain language, explains what happened with descriptive analytics, and looks ahead with predictive analytics. Every answer shows the tags and time periods behind it, so you can check it before you act.
Why plant data is hard to use
The data is trapped where it was produced.
The DCS offers OPC UA, the historian is a SQL database, the gateway publishes Sparkplug B. Core reads each of them and turns every value into the same record.
Every system describes the plant differently.
A historian’s tag tree, a SCADA’s alarm groups and a gateway’s MQTT topics each draw the plant their own way. TagLake puts one hierarchy, designed by your engineers, above all of them. Each source point maps to exactly one tag, so every value has one address.
Industrial AI projects stall on access.
Core serves live and historical values through a REST and WebSocket API described by OpenAPI and secured with roles and API keys. Each value arrives contextualised, with its unit and quality. IQ reads through that same API, so it sees what any other tool would see.
What TagLake Core does today
Everything here is in the build you can install now.
Four kinds of source
OPC UA, MQTT with Sparkplug B, Microsoft SQL Server and HTTP/JSON endpoints, all read by Core itself.
One record for every value
Each value is contextualised with its type, unit, both timestamps, quality and full source trail, whichever source it came from.
Unified Namespace (UNS)
A hierarchy your engineers design, from enterprise down to tag. Every source point binds into it explicitly.
Calculations
Write a formula once and the result becomes a tag in the namespace, like any other.
Historian and trends
History stored in InfluxDB, with a trends page of up to twelve pens. On Windows, InfluxDB is bundled.
Read-only, with roles and audit
TagLake has no way to write to your plant. Viewer, engineer and administrator roles, built to OWASP ASVS Level 2, and a journal of every configuration change.
Throughput
One figure, with the conditions it was measured under.
50,000 values/s
Sustained with no loss at 60% CPU and 153 MB. Measured on one laptop running the MQTT broker, the load generator and TagLake Core together. It is not a tuned benchmark.
461 automated tests run on Windows and Linux on every push, alongside a Windows package build and a smoke test of the portable, service and desktop modes, and a SQL Server integration job the release depends on.
Tell us what your plant runs
List the systems that hold your data and what you need from them. We will tell you plainly whether TagLake fits.