Sciento physical AI · Autonomous lab

Assayla.
Simulate it. Then run it.

A digital twin and robotic orchestration layer that converts scientific plans into validated machine steps, tests them virtually, executes them on lab automation, and records every physical action.

↳ Partner-led MVP↳ Digital twin validation↳ Immutable action trace
Assayla logo
CompilePlan to machine steps
SimulateDigital twin dry run
ExecuteRobotic orchestration
TraceEvery physical action
Why Assayla

The physical experiment is still the slowest break in the AI loop.

Sciento can design a protocol, but a person still has to translate it into instrument-specific steps, catch physical constraints, handle scheduling, and manually return the outcome. That makes execution slow, variable, and difficult to learn from.

Assayla is designed to simulate the workflow before reagents are spent, coordinate compatible automation, stream observations back through Scopea, and preserve the exact sim-to-real record.

Capabilities

A control plane for automated experiments.

01

Lab digital twin

Model instruments, deck layouts, timing, material movement, and workflow constraints.

02

Protocol compiler

Translate co-scientist plans into validated, machine-readable execution steps.

03

Robotic orchestration

Coordinate liquid handlers, plate movers, incubators, readers, and connected systems.

04

Pre-run optimization

Estimate reagent cost, scheduling conflicts, and likely failure modes before execution.

05

Closed-loop control

Use Scopea measurements to support mid-run adjustments and next-step decisions.

06

Physical audit trail

Record commands, device state, sensor output, overrides, and results for GxP evidence.

Execution loop

Close the gap between protocol and physical result.

01Compile

Make the plan executable

Resolve steps, dependencies, resources, and instrument commands.

02Twin

Dry-run the workflow

Test timing, motion, collisions, material flow, and cost.

03Run

Orchestrate the lab

Execute against connected automation with bounded policies.

04Learn

Compare sim to real

Return observations and discrepancies to the scientific loop.

Accelerated stack

Physical AI for a high-value scientific environment.

Jetson OrinRTX / H100IGXIsaac SimIsaac LabOmniverseCosmosFleet Command
Roadmap statusAssayla is a partner-led build path. The first milestone is a digital twin of one liquid-handler workflow, compiled from a Sciento protocol, simulated, then executed through a cloud-lab or automation partner.

Connect one automated workflow end to end.

Start with a bounded instrument workflow and build the sim-to-real dataset from day one.