Platform Settings
Runtime, memory, governance, card, and web engine configuration
Runtime
Max concurrent agents
Cap on simultaneously running agents
Action timeout
Max time per OAP action before replanning
Auto-replan on failure
Replanner fires automatically on verification failure
Log level
Runtime verbosity
Model Providers
Configure multiple models with distinct roles. The planner routes each task type to the right model.
Computer Access
Let your agents run tasks on your own computer by running a small local runner that you control. It executes jobs your agents queue (shell commands, screenshots) and returns results. Off by default; runs only while you keep the runner open.
Floating Core Assistant
Show the animated OpenLAM Core as a floating orb on every page, so you can summon it to chat (text or voice) from anywhere — not just the Dashboard.