AI-native simulation and emulation

Every desk can run the plant.

Transio ActionArena is an AI-native simulation and emulation platform. Describe a scenario in plain language and it runs on a sim-ready model of your plant, built from reality capture and your own CAD, with your controller logic in the loop and a signed record at the end. Every team can ask: controls, layout, material handling, quality and the launch review, each in their own terms.

Request a demonstration See what you can ask for

● reality capture · 13.3 mm median vs a survey reference● sim-ready assets from capture and CAD● your PLC program in the loop● signed evidence for every run

flow simulation · 30 s · a written request changes the AMR fleet mid-run and the two runs are compared

rendered in-house · physically based lightingsynthetic scenario

What changes

Simulation for every desk.

ActionArena is built for the people who run the plant, from the first question to the signed result. Anyone can describe a scenario, what runs is reviewable and comparable, and the model it runs on is your floor as captured, so what holds in the arena holds on the line.

Anyone states the scenarioin ordinary language: rate, belt, stations, buffers, fleets, faults, acceptance tests.

Minutes per scenariofrom templates and sim-ready assets; parts from Autodesk Inventor, SolidWorks and your other tools import ready to run.

Missing information comes back as a questionand every applied default is listed on the scenario a person approves.

Your controller logic runs in the loopStudio 5000 through Logix Echo, TIA through PLCSIM Advanced; every run ends in a signed, comparable record.

Rendered overview of a synthetic appliance line in an assembly hall

a synthetic line: infeed, vision, divert, accumulation, AMR loop, leak test, palletizing

Rendered close-up of the vision inspection station

vision inspection · a verdict tag on each carrier

the scheduled jam · alarm at TON 2000 ms · motor command dropped

The digital twin and the simulation, together

One model, from capture to commissioning.

The captured plant is the environment the scenarios run in, and every run reads back onto the twin: fit, clearances, routes, KPIs and evidence pinned where they happened. Capture again and the twin, the sim-ready assets and the scenarios move together.

Capture, from whatever fits the site
Phones and tablets, 360° cameras, handheld and terrestrial scanners, drones, fixed cameras, existing scans and BIM. We choose per site and per question, and validate to survey grade where it matters.
validated to 13.3 mm median against a survey reference 2.3 cm change floor between captures
Sim-ready assets, from any source
Every conveyor, station, fixture and vehicle becomes a sim-ready asset: measured geometry with motion and behavior. Parts from Inventor and SolidWorks, layouts from Navisworks and AutoCAD, STEP, JT and USD, supplier FMUs. Where no sim-ready asset exists, we build it for you from your CAD, the capture or the vendor's data.
measured geometry · motion · behavior
Simulate in the twin, verify on the floor
Scenarios run inside the captured environment with your controller logic in the loop; results pin to the twin; the approved line is viewed in place through an iPhone or glasses; the install is checked against the same capture.
94 % of field frames auto-localized on a stock iPhone
Captures that keep paying
The same captures and the sim-ready twin feed the plant's own AI: real imagery to evaluate vision models, labeled synthetic images from the twin to train them, camera placement rehearsed before mounting, maps and context for AMRs and assistants.
vision · safety · AMR maps · assistants

digital twin · 24 s · from the capture to the running line

1 · capture the bay and place the line

The simulated line running inside the point-cloud capture

2 · run it there, controller in the loop

3 · read results back onto the twin

A tablet showing the approved layout in place

4 · verify in place, capture again

Scenarios

What you can ask for.

Each card is a sentence an engineer might type, and what comes back.

On the line and in the cell

Throughput and takt

Run the panel line at 42 a minute. Where does it starve, where does it block?

Returns throughput, starved and blocked time, the bottleneck, WIP.

The accumulation table with carriers queued

Buffer sizing

Accumulate up to 20 before the testers. Does it absorb a five-second changeover?

Returns peak occupancy, drain time, tester starvation.

Fleet sizing — AMR / AGV

Six AMRs at 1.2 m/s at 24 units a minute — is six enough? Twelve on a 20 m loop?

Returns trips, utilisation, pickup queue.

Interlock timing and recovery

Jam at 70 s. Verify the two-second timer drops the motor and blocks restart until clear and reset.

Returns measured values against bounds; alarm at TON 2000 ms.

Leak-test booths and the palletizing cell

Test stations and changeover

Four leak testers at 4 s or six at 5 s? Twelve per pallet, five-second changeover.

Returns station utilisation, pallets completed, upstream effect.

Yield and reject routing

Seed 5 % vision failures and 3 % test failures. Is anything misrouted while faulted?

Returns the routing assertion: misrouted = 0, rejects by lane.

With your assets, controllers and data

The control panel with the fault indicators lit

Your PLC program in the loop

Run the Studio 5000 project through Logix Echo against the cell and verify the jam interlock.

Returns your logic driving the model; timing measured at the I/O boundary.

Close-up of the palletizing robot inside its guarding

Parts from Inventor and SolidWorks

Bring in the SolidWorks model of the new gripper and the Inventor carrier; check reach, cycle and footprint at the palletizer.

Returns sim-ready parts placed and run; reach and cycle re-checked; interference flagged.

Existing 3D models, made sim-ready

Import the line geometry we already have as STEP, JT or USD and give it motion, sensors and behavior.

Returns assets you can place, run and reuse across scenarios.

Layouts from Navisworks and AutoCAD

Load the Navisworks layout for the new line and place the cell where the drawing says.

Returns geometry from the layout, clearances checked, deviations flagged.

Historian data as the arrival pattern

Seed arrivals and stoppages from last quarter's historian data and rerun the fleet sizing.

Returns the scenario driven by recorded patterns, source noted on the evidence.

The pallet stack in the palletizing cell

Order data from MES

Take the next shift's orders from MES and tell me whether the palletizer keeps up with the mix.

Returns the cell loaded by the actual orders, minutes lost by cause.

On the captured floor

The line drawn over a scanned point cloud of the hall

Fit on the captured floor

Use last week's capture of bay 4 and place the panel line where the layout says. Does it fit?

Returns clearances and reach against the as-built columns and walls.

Plan view of the AMR loop over the scanned floor

AMR routes through the real aisle

Route the AMR loop through the actual aisle, 0.9 m clear of the columns and the walkway.

Returns the path on the captured floor, conflicts listed.

Scanned view of the buffer and test area

What changed since last week

Compare last week's capture with this week's. What moved, and does the layout still hold?

Returns change detection and the layout re-checked where something moved.

For the plant's AI

Captures and the sim-ready twin become data for your own models.

A labeled synthetic training set rendered from the twin

Training data from the twin

Render 10,000 labeled images of scratched and dented panels under this line's lighting for the inspection model.

Returns a labeled synthetic set from the sim-ready assets, with the real station's camera, lens and lighting.

An evaluation of the inspection model on captured frames

Evaluate a vision model on captures

Score the current inspection model against last month's captures from bay 4 and show me where it misses.

Returns precision and recall by defect type; misses pinned to the twin by station and shift.

The vision station with its camera field of view

Camera placement, rehearsed

Where should the second camera go to see the whole reject lane without glare from the skylight?

Returns field of view, occlusion and lighting checked in the twin before anything is mounted.

And in augmented reality

The same model can be viewed in place. Our field AR localizes a stock iPhone or iPad from what the camera sees, marker-free, and overlays the approved layout on the real floor; glasses carry work instructions and capture observations hands-free.

A tablet showing the approved layout standing in the bay

How a request becomes a run

Describe, review, approve, run, compare.

01 · Describe

The request, in ordinary language.

Rate, belt, stations, buffers, fleets, faults and acceptance tests, in any order. Anything under-specified comes back as a question.

compose

02 · Review and approve

The scenario, exactly as it will run.

Layout, I/O map, timing, acceptance tests and applied defaults, laid out for review. A person approves it; any change comes back for approval.

review · scenario · approval

03 · Run and compare

Your controller logic in the loop; the record signed.

Scheduled faults fire; results are checked against acceptance bounds; the run is recorded and signed. Approved scenarios compare like-for-like.

evidence · commissioning checks

controller emulation · 26 s

obstruction, timer, alarm, motor command dropped, restart blocked until clear, reset and start

Assertion	Measured	Bound	Result
jam stop within bound	1999.17 ms	1985 – 2040 ms	pass
alarm at timer preset	TON 2000 ms	= 2000 ms	pass
motor command dropped	4.8 ms	≤ 20 ms	pass
restart blocked until cleared	held	—	pass

from a completed, valid run

Ask the twin, ask the run

The plant and its runs, questioned in plain language.

The same layer that takes a scenario also answers questions about the captured plant and about completed runs, with the source of each answer attached.

The palletizing cell with a clearance measurement pinned

1.42 m · fence to column

Layout engineerHow much clearance is there between the palletizer fence and column C3?

ActionArena1.42 m at the closest point, fence post 2 to the column face. The walkway line runs 0.9 m from the fence, so the walkway holds. Measured against the 12 Aug capture.

capture 2026-08-12layout rev C

Scanned view with a moved station marked

Leak test 2 · +8 cm east

Industrial engineerWhat moved since last week's capture, and does the layout still hold?

ActionArenaLeak test 2 moved 8 cm east; everything else is within the 2.3 cm change floor. The AMR loop clears it by 0.71 m; the layout holds.

capture 08-05 vs 08-12change detection

Material handlingIn run B, when did the accumulation table last hit eight, and how long was the line blocked?

ActionArenaFull from 142 s to 171 s of process time. Upstream blocked for 7.9 s in that window; after the fleet change to ten, blocked time is 0.0 s and the table drains to two.

run B · evidence record

Launch reviewCompare A and B on pickups per minute, blocked time and tester utilisation.

ActionArenaBoth runs are the same approved scenario with one change, the fleet size, so the comparison is like-for-like.

metric	A · 4	B · 10	Δ
picked up / min	25	47	+22
line blocked	7.9 s	0.0 s	−7.9 s
tester utilisation	31 %	49 %	+18

illustrative exchanges on the synthetic scenario

Interactive model

The same cell, in your browser.

Orbit the line, schedule a jam and recover it, change the fleet or the arrival rate, switch the environment between the model and the capture, and watch the controller I/O. It runs entirely on this page.

Requires WebGL. Drag to orbit, scroll to zoom, shift-drag to pan. Switch the environment to the capture to see the same run inside the twin.

Scenario calculator

Arrivals30 /min

Vision reject rate6 %

Belt speed0.60 m/s

Leak testers4

Test dwell4.0 s

AMR fleet4

AMR speed1.5 m/s

Robot pick cycle1.4 s

Pallet changeover5.0 s

Capacity by stage against demand

Who it is for

Every team that touches a line.

Controls and automation

Rehearse interlocks, timers and recovery sequences with a controller in the loop, and keep the evidence.

Industrial engineering

Size buffers, testers and changeovers against the real mix, and compare options like-for-like.

Material handling

Size and route AMR and AGV fleets on the actual floor, with queue and blocked time measured.

Quality

Seed defect rates and check routing, holds and rejects under fault conditions.

Launch and program management

Compare designs, sequence retools and check fit on the captured floor before steel moves.

Maintenance and OT

Exercise network resilience and recovery procedures on the simulated cell, with production untouched.

Contact

Request a demonstration.

Tell us about a line, a cell or a question you would like to rehearse. We will reply within two business days.

The form opens your mail client and the message comes straight to us.

Transio, Inc.

2261 Market Street, STE 85661, San Francisco, CA 94114
gnain@transio.ai
ActionArena · Twin · Field AR · Cosmo

How engagements start

Discovery conversation to choose a workflow and a sandbox.
Four-week demonstration on that workflow.
Pilot on one line, one interlock set, your PLC program in the loop.

Videos, renders and the interactive model on this page were produced in-house from a scripted scenario studio and are illustrative. Product screens are captures of the running prototype on synthetic data. Controller and CAD integrations are qualified per engagement.