Automatic twins
Turn CAD or a scan into a simulation-ready site in hours.
Features
Everything to train and run a fleet
From twin construction to edge inference, Foremai gives your team a complete toolkit for physical AI.
Deploys on your robots Validated in NVIDIA simulation No hardware at risk
Core
The features teams rely on
A focused toolkit for building, validating, and orchestrating skills.
Parallel training
Thousands of GPU environments training at once.
Safety gates
Success, collision, and force checks before deploy.
Agentic planner
Goal decomposition and live fleet coordination.
Edge inference
Real-time policy and perception on Jetson.
Ops analytics
Throughput, uptime, and cost-per-task dashboards.
Simulation
Train thousands of robots at once
Isaac Lab spins up massively parallel environments so policies converge fast and cheaply.
- RL and imitation learning
- Domain randomization out of the box
- Reproducible, versioned runs
4,096 parallel envs
GPU-accelerated
Sim-to-real report
Versioned checkpoints
Developer-first
Train a skill in a few lines
The Fore SDK makes training and validating a policy feel like shipping software.
train_palletize.py
from foremai import Studio, Twin
twin = Twin.load("northforge/line-3")
skill = Studio.train(
task="palletize_mixed",
twin=twin,
envs=4096,
gates=["success>0.98", "no_collision"],
)
skill.validate().publish("fore-market") # signed + safety-gatedfore deploybash
$ fore deploy palletize_mixed --site line-3
✓ safety gates passed (success 0.993, 0 collisions)
$
→ publishing to 42 Jetson agents...
$
✓ live on line-3 · fleet coordinating
Deploy
Ship it with one command
When a skill clears its gates, deploy it to the fleet from the CLI.
Orchestration
How the agent thinks
The planner turns a shift goal into coordinated robot action.
1
Understand
Parse the shift goal and constraints.
2
Decompose
Break it into robot-executable tasks.
3
Schedule
Assign and sequence across the fleet.
4
Adapt
Re-plan on exceptions in real time.
Performance
Fast where it counts
12ms
edge control loop
4096
parallel envs
99.3%
skill success
6x
tasks per robot
Integrations
Works with the stack on your floor
Vendor-agnostic across robots, controllers, and systems of record.
Robots & arms
Universal RobotsFANUCKUKAABBFetchBoston Dynamics
AMRs & humanoids
MiROTTOLocusAgilityApptronikFigure
Controllers & sensors
ROS 2PLCs / OPC-UAIntel RealSenseZividSICKCognex
Systems of record
SAPManhattan WMSBlue YonderNetSuiteSnowflakeDatabricks
Security & safety
Enterprise-grade, safety-first by design
Robots act in the physical world, so every skill is validated before it ever touches hardware.
SOC 2 Type II ISO 27001 GDPR ISO 10218 / RIA R15.06 SSO / SAML Audit logs
Safety gates
Every policy must pass success + collision + force limits in sim before deploy.
Encryption everywhere
AES-256 at rest, TLS 1.3 in transit, per-site key isolation.
Data residency
Run in your VPC or on-prem; telemetry never leaves your boundary.
SSO, SCIM & RBAC
SAML/OIDC, provisioning, and role-based access down to the cell.
Feature FAQ
Questions about features
Domain randomization plus telemetry-driven re-calibration keep the twin tight. Every skill ships with a sim-to-real report so you can see the gap before deploying.
Yes. Fore Fleet supports human-in-the-loop approvals, teleop handoff, and hard stops at any time.
Model the end-effector once in the twin and Foremai trains against it. New tools are a configuration, not a rewrite.
Get started
See the features in action
Book a demo and we’ll train and deploy a skill live.