Platform

Shape · field · state · plant

The engines underneath.

Simulation fast enough to control against.

Acaysia's controller samples thousands of futures every tick and steers along their cost-weighted path. The models behind each reactor are trained on AcaysiaRT, AcaysiaDRT and Rete, so the controller looks ahead before it moves a setpoint, and hands back to your PID in under 100 ms.

  1. 01Shape

    AcaysiaGEM

    The vessel, fit to simulate

  2. 02Field

    AcaysiaDRT

    Every cell resolved in 3D

  3. 03State

    AcaysiaRT

    Lumped state, stepped in time

  4. 04Plant

    Rete

    One unit in a plant with recycle

Schematic — one reactor, four resolutions: geometry, 3D transport field, lumped state, flowsheet unit

How it fits

One API · one job format

Four engines, one surface.

Each engine answers a different question about the same plant. All four speak one contract family, and AcaysiaCORE serves them from one surface.

GEM builds the vessel, DRT resolves it in 3D on top of AcaysiaRT, and Rete wires AcaysiaRT models into whole plants. AcaysiaCORE brings all four into a single browser application.

Schematic — four engine domains converge through one contract family into one AcaysiaCORE surface, served to a single browser

AcaysiaRT

Throughput

Reactor-steps per second.

On one RTX 5080. AcaysiaRT is the runtime underneath the Acaysia MPPI control system — fast enough to sample thousands of futures every control tick.

AcaysiaRT

The Acaysia Control System

PID fallback < 100 ms

See Acaysia in action.

Start with a discovery call about your process, or ask for early access to AcaysiaCORE and run the engines from your browser.

Book a discovery call Request CORE early access