ADMINISTRATOR ACCESS
Your infrastructure.
Your AI workspace.
Authorize this browser from your existing administrator SSH session. Browser access permits inference and read-only server diagnostics. Production changes retain the existing console approval workflow.
AI CONTROL CENTER
What are we working on?
Ask for analysis, documentation or coding guidance. The local model has no shell or privileged server access.
PROVIDERS
Agent availability
EXECUTION HISTORY
AI Task Center
EXISTING MANAGED INVENTORY
Server operations
Read-only diagnostics use the existing restricted node agents. Restarts and deployments require the existing human console approval workflow.
EXECUTION POLICY
AI settings
One active inference job · two CPU threads · 120 second deadline
Local model: Qwen2.5-Coder 1.5B · 4,096 token context · 512 output tokens
Protected task history: 30 days · 20 queued jobs per operator · 2,000 retained tasks maximum
Automatic routing uses authorized adapters, then local inference. Paid fallback is disabled. Provider quotas remain unknown until authorized authentication.
Production operations: external approval required. No arbitrary shell execution.