Request a Demo

Quality can't be tested in. It has to be measured on the way through.

The lab tells you what happened five days ago. It doesn't tell you when to stop. Newton reads concentration inside the column while the run is still going — so the call is made on a number, and every run behind it is a record you can repeat.

Measure the process, not just the output

Concentration read live in-line, while the run is happening, instead of five days later from a lab.

One record, not five systems

Sensor data, operator decisions, and lab results connected to one batch that follows the material to finished product.

An end point, not a timer

The run ends when the material hits the number, not when the clock does. Every other manufacturing sector already made that shift. Extraction is making it now.

Built around the process-understanding framework FDA established for regulated manufacturing.

Already deployed in multiple facilities
LeefSungrownAbstraxNimbusEuphoric Life
Two Runs, One SOP

Every extraction is unique.
Your SOP isn't.

Your starting material is a plant. No two lots run the same, so standardizing extraction takes a tool that flexes with the run instead of a fixed time and a judgment call.

Run A Run BTHC concentration, read live in-line
0%2%4%6%8%10%SOP stop · same time, every runA: spent 10 minutes agoB: still producingstartA's windowB's windowend of run

One SOP. Two different mistakes.

Where you stop inside the window is your call. It depends on what you're making. Newton doesn't make the decision, it shows you the window while you're still in it.

Here's what the top teams already know

The difference isn't the variables. It's the platform.

Common Questions

A few things people ask first.

Yes. It bolts inline on your existing setup with no equipment to replace, and it works with hydrocarbon and ethanol extraction today.

Yes. The sensor that goes inline is rated for the classified area, and all electronics sit outside the booth, so nothing uncertified is in your C1D1 space.

The readings are accurate to within about 2%, similar to an analytical testing device. Like any inline device, accuracy depends on a clean window, and cleaning it is the same process as cleaning your skid. Your normal skid cleaning schedule covers it in most cases.

The sensor and your run are independent from the data layer, so neither a dropped connection nor a hardware fault can alter your extraction system.

Thirty minutes with someone who knows extraction. We walk through a real run — the curve, the peak, and where the stop got called — and answer whatever you want to ask. If it's not a fit, we'll say so and that's the end of it.

Pricing is per site. You can buy the system outright or lease to own, whichever fits your budget and setup.

See where your peak actually is.

Thirty minutes, on a real run. A reply within one business day, from a person.