SCDI · Supply Chain Disruption Investigator
Agent Offline
● Ingesting Open-Meteo, USGS, GDACS & Wire Feeds

Real-time disruption & predictive risk radar

Autonomous Planner ↔ Executor ↔ Evaluator loop driven by live multi-API global weather, natural disaster feeds, and geopolitical current events. Evaluates predictive disruption likelihood and supply chain impact.

🔄

Contacting Open-Meteo, USGS, GDACS and global intelligence feeds…

Global Supply Chain Route Risk MapReal-Time Node Dynamic Telemetry
ShanghaiNingboBusanKaohsiungSingaporeSuezJebel AliRotterdamAntwerpHamburgLos AngelesLong BeachHoustonSavannahBaltimoreNew YorkPanama CanalCape Bypass
Executive Disruption Impact & Value RealizedAgent Mitigation Verified
Decision confidence: 94%
Total Cost Avoidance▲81% Saved
$2.71M
Avoided penalty & spot surgeRisk: 85%
Days of Inventory (DOIR)+8.4d Restored
19.5 Days
Safety Buffer
Buffer threshold: 14.0 daysHealthy
Protected On-Time DeliveryDelay: +14.5d
68%
33% raw
PO split & dynamic reroute+35% uplift
Agent Decision LatencyO(1) Flat Context
340 ms
NVIDIA NIM
180 tokens per chunk3 corridors rerouted
Autonomous PEE Graph Engine (Planner ↔ Executor ↔ Evaluator)Cycle status: idle
DAG steps
Planner
Zetaris / tools
Executor
NIM scoring
Evaluator
Refine / Retry feedback
Bounded Markovian Chunks & Carryover Memory3 Iteration Chunks · $O(1)$ Flat Context
Chunk 1 · Markovian
RETRY
Historical high shock → DOIR 0.0d; missing backup contracts + alt lane capacity
DOIR 0d · burn 860/d
Chunk 2 · Markovian
CONTINUE
Backup V-482 via rail_intermodal recovers DOIR to 4.7 days (hist delay 11.8d)
DOIR 4.7d · burn 860/d
Chunk 3 · Markovian
PROCEED_TO_FINAL_OUTPUT
Mitigation accepted for Route stress R-PAC-USW: Shenzhen → Los Angeles: reallocate PO to V-482
DOIR 4.7d · burn 860/d

Carryover state (into next chunk)

Waiting to run…

Structured JSON trace

{}