AI that prevents downtime in autonomous logistics

Pulsar turns ports, yards, and warehouses do into simulation environments where teams can handle failures before they happen.

Ports
Gaussian splat reconstruction
Warehouses
Point cloud segmentation
Yards
3D mesh reconstruction

01
One image
A single photograph of the terminal is all we need. No LiDAR, no ground truth survey.
02
Scene reconstruction
Monocular depth + gaussian splatting build a dense 3D representation of the static environment.
03
Asset placement
Cranes, AGVs and vessels are detected, parameterised and placed as controllable rigid bodies.
04
Failure rehearsal
Operators run physics-accurate failure scenarios in the twin, before any real incident occurs.
Explore our pilots
Original Backdrop Simulation
Already mapped & test ready
Pending data acquisition
Upcoming targets
$14T
Global port trade
800+
Container ports
< 5 min
Image to simulation
94%+
vs. LiDAR accuracy

Built by engineers who've seen the failures firsthand

We are a team of robotics engineers, simulation researchers, and port operations specialists. Pulsar was born from a simple observation: the terminals most exposed to automation risk are also the ones least equipped to rehearse failure.

Our pipeline turns one photograph into a physics-accurate digital twin in minutes, no LiDAR survey, no proprietary data format. From that twin, operations teams run the scenarios that matter: crane collisions, AGV cascades, power-loss emergencies. Practice the worst before it ever happens on the apron.

Pulsar team at the port
COMING FROM Hugging Face Partner Partner Marso Robotics MIT
Generated World
Powered by URDF Studio Open in new tab ↗

URDF Studio cannot be embedded due to browser security policy.

Open URDF Studio ↗ Your generated URDF has been prepared and is ready to import.
Open URDF Studio ↗

Import scene_rig.urdf to run physics simulation.