medOS AI
AI Command Center
Every model, every surface, one pane
All LLM and AI surfaces grouped in one operations console — usage from the audit log, per-tenant flags, runtime fabric across edge boxes, on-prem servers and cloud, with PHI residency enforced at the routing layer.
Edge box
qwen2.5:3b
8 nodes
On-prem
mistral:7b
2 GPUs
Cloud
claude-haiku
PHI-safe only
voice-order412 calls
smart-diagnosis238 calls
coder-suggest1,104 calls
intake-classify97 calls
PHI never falls back to cloudper-tenant kill switches
Capabilities
Live usage panel over the per-inference audit log
AI runtime fabric: edge box / server / cloud tiers with heartbeats
PHI-heavy inference never falls back to cloud — queues or degrades to manual
Per-tenant kill switches for every AI surface