Stop building simulation. Start running it.

Teams building autonomous forklifts, pallet movers, tuggers and goods-to-person AMRs. Autonomy and simulation engineers keep the planner, the perception stack and the corner cases, and hand Forge the scene building, asset wrangling, sensor modelling and scenario scaffolding. Test, QA and release teams run the full suite on every release candidate. Field engineering and sales rebuild a prospect's site and run the robot through it before the pilot.

Forge does not replace your simulation engineers; it takes the work that does not need them. The scene, the assets, the sensor models, the scenario scaffolding and the first pass at root cause are drafted by Forge. Your engineers review what it proposes and stay on the problems only they can solve.

Your robot's URDF and one site that gave you trouble. Forge imports and tunes the robot model, places and calibrates the sensors, and builds the site; your engineers check it against the real thing before anything runs.

No. Forge models the vehicle and its sensors; your nav2 and ROS 2 stack connects and drives exactly as it ships. The build that drives in the field drives in the twin, so a pass in Forge is a pass for your release, not for a simulation-friendly variant.

No. Forge evaluates the policy as a black box over a secure link. Model weights, source code and proprietary logic stay inside your environment; Forge only exchanges the sensor inputs and control outputs needed to run the scenario.

No. The drive model and the sensor suite change; the process does not. A counterbalance forklift, a tugger in a plant, a pallet mover at a dock and a goods-to-person AMR all go through the same build, draft, run and grade loop.

From a description or a walkthrough video. Forge lays out racking, docks, aisles, lanes and zones, names each area so scenarios can refer to it, and flags anything it is unsure about. Your engineers correct whatever does not match the real site.

The six conditions that decide a deployment: people stepping out in shared aisles and doorways, dynamic obstacles left in the lane, narrow passages and precise alignment, localization under stress, degraded perception, and mixed traffic. Forge drafts each one for the site in question and proposes the variations worth running, so the suite covers more than the cases someone remembered to write.

The measurement that decided it, with its limit and its signed margin. The route the robot actually took, ground truth from the simulation, against what nav2 planned and what the robot believed about its own position. The exact conditions the run drew, so it can be replayed deterministically. And a root cause traced to your own rosbag data: which message, which timestamp, which decision.

The full suite runs on every release candidate and the report gives a release-over-release diff: what held, what broke and what was fixed. Failures that recur across runs and sites are reported as trends, so a quiet regression does not wait for the field to find it.

The Safety suite references ISO 3691-4 and ANSI B56.5, and covers people at aisle ends, doorways and crossings, speed zones and slow-downs, emergency stop and protective-field switching, and blind corners. Each result is traceable to the requirement its scenario was written against.

Yes. The Integration suite covers right of way, crossings and deadlock with other vendors' robots and manual equipment, and VDA 5050 task handoff from a WMS or fleet manager.