See the batch while it runs.
Biomass, substrate and specific rates arrive hours late. We turn hard-to-measure process state into timely decisions, so operators act before the window closes.
Bioprocess and bioreactor control
Stop flying blind. Keep every batch in control.
Labinetix runs estimation and control models on your industrial controller, so the state you cannot measure is visible while the batch runs, and every run behaves like the last.
Start with one batch How it works
- Fed-batch and bioreactor processes
- Runs on the controller, not in the cloud
- Deterministic by design
Built for teams who cannot wait on the lab.
Living processes move faster than the information about them. Labinetix closes that gap while the batch can still be steered.
Keep automation consistent.
Hand-tuned logic drifts between runs. We deliver deterministic control teams can trust from inoculation to harvest, with fewer end-of-batch surprises.
From signal to control, one path.
We close the loop between what the plant measures and what the controller does, without a detour through a spreadsheet or a second model of the truth.
-
Sense.
Start from signals the plant already produces: off-gas, dissolved oxygen, pH, feed, agitation, at controller cadence, not lab turnaround.
-
Estimate.
Filtering, rate estimation and elemental balancing turn those signals into growth, uptake and production, kept physically consistent by closure.
-
Control.
Deterministic control executes on the industrial controller itself: fixed cycle, bounded memory, no allocation in the critical path.
-
Prove.
Every model, interface and parameter set is a versioned artifact, and every deployment and every run is recorded, so what you qualified is provably what ran.
Start with one batch.
A pilot is one process, one controller, one campaign.
We look at your process and your controller.
What you cannot measure in time, what you already measure, and what runs the loop today.
We build and run it beside your automation.
The estimator and the control for that one process run on your controller for one campaign, next to what you already trust.
You keep the result.
The software stays on your controller, with a report of what was predicted against what happened.
Engineered for the controller, not the cloud.
Real-time control on a plant is unforgiving. Our stack is built for that constraint, so model and plant stay aligned where it matters.
How the platform works- Deterministic numerical core. Control and estimation kernels written for the controller: no allocation on the hot path, no dependency on a runtime that is not there.
- Runs on your controller. IEC 61131 controllers with a Linux runtime, and Linux-based industrial edge controllers. Tell us which one you run.
- One path from model to plant. The definition that drives simulation is the definition that generates what ships, with no reimplementation step in between.
Tell us about your process.
Whether you are fighting blind batches, inconsistent automation, or a model that never made it to the plant, a short conversation is usually enough to tell whether we can help.