Key facts
| Access control | Scoped API keys with rotation; enterprise SSO and RBAC options |
| Partner access | Separate keys per carrier, broker or customer integration |
| Deployment | Cloud, VPC, on-prem or air-gapped across regions and hubs |
| Auditability | Request, model and response logging for dispatch and dispute review |
| Data grounding | Embeddings and RAG are live for tariffs, lanes and customs guidance |
| Models | 30+ models behind one OpenAI-compatible API |
| Pricing model | Flat monthly self-serve plans; no per-token billing on self-serve |
| Endpoint roadmap | Files, batch and fine-tuning are coming soon |
TL;DR
- Classify shipment, customer, customs and contract data separately.
- Give each carrier, broker and customer integration its own scoped key.
- Document residency for cross-border lanes before automating them.
- Log dispatch decisions so disputes and incidents can be reconstructed.
- Route exceptions, claims and reroutes to named human operators.
How it works, step by step
- Inventory AI use cases across quoting, dispatch, tracking, customs and claims.
- Classify data by customer, lane, jurisdiction and contractual restriction.
- Choose deployment per class: region-selected cloud, VPC, on-prem or air-gapped.
- Issue per-partner and per-application keys with rotation and central inventory.
- Define log fields and retention: request ID, model, sources, output, operator decision.
- Approve a model allow-list and ground answers in current tariffs and lane rules.
- Define which exceptions must escalate to a human operator and how.
Try it yourself
Open the AI workload router simulator →
Shipment, customer and contract data
Logistics data includes addresses and contact details, shipment contents, customs declarations, rate contracts and customer SLAs. Each class needs its own handling rule, and cross-border lanes add jurisdiction-specific restrictions on what may be processed where.
Most teams start with public tariff and customs guidance, then extend to customer shipment data only on private deployments where prompts, documents and embeddings stay inside the environment.
Keys, partners and least privilege
Issue a distinct API key per carrier, broker, customer integration and internal application. Store keys in a secrets manager, rotate on a schedule, and revoke them when a contract ends or a partner is replaced, which happens frequently in logistics. Enterprise SSO and RBAC options keep console access aligned with depot and operations staff turnover.
Never share a production key across partners, and keep customer identifiers out of prompts where retrieval can supply only the fields a task needs.
Residency, retention and dispatch audit trails
Decide where processing happens for each lane and how long records live. Region selection covers many needs; VPC, on-prem and air-gapped deployment covers networks that require data to stay in a country or inside the operator's estate. Retention applies to prompts, outputs, logs and retrieval indexes, and customs records follow their own schedules.
Log enough to reconstruct a dispatch decision: request ID, model and version, retrieved source identifiers, output, operator and lane. See AI audit logs for a schema.
Model governance and exception handling
Keep an approved model list with evaluation evidence and re-test when versions change. Ground answers in current tariffs and lane rules with citations so operators can verify during a live exception. Claims, reroutes and compensation stay with named humans, and the checklist should state which decisions may be automated and which must escalate.
Honest comparison
| Control area | Plugsky capability | Common gap | Owner |
|---|---|---|---|
| Partner access | Separate scoped keys per carrier and integration | Shared portal credentials | Integration engineering |
| Data boundary | Region choice plus VPC, on-prem or air-gapped deployment | Lane data path undocumented | Compliance |
| Retention | Configurable logging of requests and responses | No defined retention | Records management |
| Audit trail | Request, model, lane and operator logging | Disputes not traceable | Operations |
| Grounding | Embeddings and RAG over tariffs and customs rules | Outdated rate answers | Pricing team |
| Escalation | Structured output for operator review | Automated reroute promises | Network operations |
Frequently asked questions
Does using Plugsky make us compliant?
No. Compliance is your program. Plugsky provides deployable controls - scoped keys, private deployment, logging - that you document and audit against your own contractual and regulatory obligations.
Can we keep lane data in a specific region?
Yes. Region selection covers many requirements, and VPC, on-prem and air-gapped deployment keeps data inside a chosen environment or country.
What should we log for dispatch decisions?
Request IDs, model names and versions, retrieved sources, outputs, the operator and the lane, retained under your policy so disputes and incidents can be reconstructed.
How do we handle carrier integrations securely?
Issue one scoped key per carrier, rotate on a schedule, and revoke immediately when a contract ends. Never let partners share a production key.
Can we automate customs documentation?
Retrieval over current customs guidance works today with citations. Keep a human review step for declarations, and note that file and batch endpoints are coming soon.
Is fine-tuning available for our rate logic?
Fine-tuning and batch endpoints are coming soon. Today, use retrieval over approved rate cards with JSON mode for consistent structured output.
Where should a pilot start?
Pilot on public tariffs and customs guidance with the free plan, measure answer quality, then extend to customer shipment workflows on private deployment.