One benchtop unit. Everything a fermentation needs.
Vessel, probes, pumps and interface in one connected unit. Scroll to take it apart.
The bioreactor with a process engineer built in.
A benchtop bioreactor, 1–10 L, that watches every run, records every change and helps you work out what happened. So the next experiment starts from data, not from a hunch.
It proposes. You decide.
Book a live demo30 minutes: the hardware, the interface and the agent, running.
Most bioreactors control a run. Very few help you understand it.
Day four of a fed-batch. Something drifted overnight. By morning the curve looks different and nobody saw it happen: you have pH and temperature, not what changed, when, or who touched it.
Dissolved oxygen · day 4 Example data
08:30, first look. The curve fell at 02:30 and nothing in the record says why. The day starts with a guess.
- Feed rate 12 → 20 mL/h · changed by the night operator · reason: “ramp step 3”
- Dissolved oxygen below 30 % · low-DO alarm, on the record
- You start from the cause, not from a guess.
It's the same story when a process that worked at the bench behaves differently at the next scale: without a record, the meeting ends with opinions, not answers.
1–10 L of glass and steel.
A borosilicate vessel on a stainless stand. A heating blanket and a cold finger hold the temperature; a brushless drive turns the impellers.
Every probe on the lid. Every line in one arc.
Digital pH and dissolved-oxygen probes and a PT100 sit on the headplate. The motor cable plugs in on top of the drive; probe cables and feed lines travel with it from the tower.
Four pumps. Four bottles. One path to the lid.
Acid, base, feed and antifoam. Each silicone line rises from its bottle, runs through its own peristaltic pump and follows the arc into the vessel.
Weighed, not guessed.
A load cell for gravimetric feeding and mass balance. Air and oxygen through mass flow controllers. The control loops run here, at the bench.
The interface lives on the tower.
A screen on the reactor, and the same interface on your laptop, tablet or phone, in the lab or at home.
This is what the agent reads.
Every reading, every change, on the record. From here, the agent takes over.
It watches every run. It proposes. You decide.
- BeforePlan it properly.
Set up the run with the agent: objective, variables, what you expect to see. It keeps that context for the rest of the run.
- DuringSomeone is always watching.
The agent reads the live state of the reactor and flags what's drifting, with an explanation you can check against the data. At 3 a.m. too.
- AfterUnderstand what happened.
Ask about the run in plain language. Get the plots, the events and the changes in one place, and a record of the batch you can hand to someone else.
On your phone, wherever you are. The assistant reads the live run and explains it. It doesn't touch the reactor.
Yes, “AI in bioprocessing” sounds like hype. Here's what ours does and doesn't do.
- It never changes your process on its own.It proposes a setpoint, a step, an experiment. A person approves it.
- Everything it suggests is on the record.What it proposed, who accepted it, when and why.
- It doesn't replace your technician.It replaces the alarm nobody sees at night and the afternoon spent copying data into a spreadsheet.
- Control stays at the bench.The reactor runs its loops locally. If the network goes down, the run doesn't.
Interface
One interface. Every screen in your lab.
It runs in the browser: on the reactor's own screen, a laptop, a tablet or your phone, in the lab or at home. A training mode with a simulated reactor lets you learn without risking a run.
- Live
Readings, setpoints and every actuator at a glance, with the state of each loop.
- Trends
pH, temperature and dissolved oxygen over the whole run, with your annotations on the curve.
- Recipes
Phases with entry conditions: the run moves on when the culture is ready, not when the clock says so.
Measures
- pH · industrial digital probe
- Dissolved oxygen · industrial digital probe
- Temperature · PT100
- Weight · load cell, for gravimetric feeding and mass balance
Acts
- Four peristaltic pumps · acid, base, feed, antifoam
- Heating blanket and cold finger
- Gas through mass flow controllers · air and oxygen
- Agitation · brushless drive
Runs on
- A browser interface · on the reactor's screen, or on your laptop, tablet or phone, in the lab or at home
- A training mode with a simulated reactor · to learn and test without risking a run
Working volume
- 1–10 L range
Built to be trusted
The safety and traceability most benchtop systems leave out.
- SafetyEmergency stop
Puts every pump, heater and valve in a safe state. Anyone can press it.
- SafetyWatchdog in the controller
If the computer stops responding, the reactor stops pumping and heating by itself.
- DataAudit trail that can't be quietly edited
Append-only and chained: who changed what, old and new value, result and reason.
- DataUsers and roles
Operators run, supervisors change setpoints.
- DataYour data, exportable
Every run, every variable, every event, as CSV.
Designed with the data integrity expected in regulated environments in mind.
Team
When you call, you talk to the people who designed it.
Bioprocess engineer. Has taken fermentation processes from the lab to GMP plants, led full-plant site acceptance tests (one of them in Canada), designed a single-use bioreactor from scratch and led the integration of AI into bioreactor control.
Fifteen years at the bench and in the plant. Has started up a GMP pilot plant and run upstream from 3 L to 200 L, with bacteria and plant cells. Makes sure the reactor arrives tuned, not as a three-week project.
Mechanical engineer with a background in pharmaceutical equipment and bioreactor design. Designs out the failures he's already seen in the field.
Who it's for
Built for teams developing a fermentation process that has to work at the next scale.
A good fit if you
- develop microbial fermentation processes at bench scale
- need to understand, not just control, what happens in each run
- want every change on record, ready to share with your team, your investors or your QA
Not the right fit (yet) if you
- need a GMP-certified system for production today
- run mammalian cell culture that needs CO₂ for pH control
Questions
Questions you're probably asking
You're a new company. Why should I trust you?
The company is new. Our time in the plant isn't. Book a demo and judge the equipment and the people for yourself.
Will the AI change my process without asking?
No. It proposes; a person approves. Every suggestion and every decision is logged.
What happens if the network goes down?
The run continues. Control lives in the reactor, not in the cloud.
Can I check on a run from home?
Yes. Readings, alarms and the assistant are on your phone wherever you are. Control stays at the bench: the reactor runs its loops locally.
Can I use it in a regulated environment?
It's designed with data integrity in mind: roles, an audit trail that can't be quietly edited, full export. It isn't a validated GMP system out of the box. Tell us what your QA needs and we'll tell you honestly how close we are.
How much does it cost?
It depends on configuration. We'll give you a figure on the first call.
Book a demo
See it run.
- 1Pick a slot.30 minutes, online.
- 2Tell us about your process.Organism, scale, what's been hard so far.
- 3Watch it live.The hardware, the interface and the agent.
- 4Get a straight answer.Whether it fits your process, and if not, why.
Suppliers & partners
We source vessels, heating, probes, pumps and drive components from a small number of manufacturers we work closely with. Quotations, samples and new components: [email protected]