Feature
Simulations
Simulations let teams test an agent against a place-like environment before using it in the field.
In the current product, a simulation can connect an agent to one or more environments, places, splat assets, starting poses, movement settings, and trigger context. When you run an agent from the web app, the simulation is one of the inputs the agent can look through; the other is your device camera. See Running an agent. Runs use the same server-owned session path as the phone and glasses clients.
Simulations are useful for checking:
- whether the agent understands the selected place and workflow
- whether instructions, files, and tools are enough for the task
- how the agent behaves when visual context changes
- whether trigger rules or spatial context need adjustment
Simulation does not replace field testing. It makes field testing more focused by catching basic behavior and context issues earlier.