Key facts
| Access control | Scoped API keys with rotation; enterprise SSO and RBAC options |
| Deployment | On-prem, air-gapped, VPC or edge placement at field sites |
| Auditability | Request, model and response logging for safety and ops review |
| Data grounding | Embeddings and RAG are live for procedures and equipment knowledge |
| Structured output | JSON mode returns inspection and incident records in fixed schemas |
| Models | 30+ models behind one OpenAI-compatible API |
| Pricing model | Flat monthly self-serve plans; no per-token billing on self-serve |
| Endpoint roadmap | Vision, files, batch and fine-tuning are coming soon |
TL;DR
- Classify seismic, reservoir and JV data as your most sensitive content.
- Design for remote sites: on-prem, edge or air-gapped placement.
- Give each contractor and asset team its own scoped key with expiry.
- Log safety-related requests and outputs so incidents are reconstructable.
- Keep operations and safety decisions with qualified staff.
How it works, step by step
- Inventory AI use cases across maintenance, safety, drilling support and logistics.
- Classify data by commercial value, JV rights, safety relevance and site confidentiality.
- Choose placement per class: region cloud, VPC, on-prem, edge or air-gapped.
- Issue per-asset and per-contractor keys with rotation and expiry dates.
- Define log fields and retention: request ID, model, sources, output, reviewing engineer.
- Approve a model allow-list with evaluation evidence and re-test on version changes.
- Require qualified review before output informs operations or safety procedures.
Try it yourself
Open the private LLM deployment estimator →
Field, geological and contract data
Oil and gas content spans public technical standards, internal procedures, equipment data, seismic and reservoir interpretation, and joint-venture records. The last three are commercially sensitive, governed by partner agreements, and effectively national-asset grade in some jurisdictions.
Start with public standards and standard procedures, then extend to subsurface and JV content only on deployments inside your own boundary with retrieval indexes you control.
Keys, contractors and least privilege
Issue a distinct API key per asset, contractor and application, with expiry dates that match contract terms. Store keys in a secrets manager, rotate on a schedule, and revoke on demobilisation. Enterprise SSO and RBAC options keep console access aligned with rotation schedules and crew changes, which are frequent and often remote.
Never place partner data or well identifiers in prompts where retrieval can supply only the fields a task requires.
Remote-site residency and safety records
Decide where inference runs for each workflow. Region selection covers corporate use cases; on-prem, edge and air-gapped deployment covers field workflows where connectivity is limited or data must stay in a country or inside the operating company. Retention applies to prompts, outputs, logs and retrieval indexes, and safety and incident records follow their own schedules.
Log enough to reconstruct an output: request ID, model and version, retrieved source identifiers, output and the reviewing engineer. See AI audit logs for a schema.
Model governance and operations review
Keep an approved model list with evaluation evidence and re-test when versions change. Ground answers in current operating procedures and equipment documentation with citations so staff can verify during an upset condition. Safety-critical guidance stays with qualified personnel, and the checklist should name the reviewer for each workflow and define what may never be automated. Vision endpoints are coming soon; plan inspection workflows around that status.
Honest comparison
| Control area | Plugsky capability | Common gap | Owner |
|---|---|---|---|
| Placement | On-prem, edge, VPC or air-gapped deployment | Subsurface data sent to public tools | Operations IT |
| Identity | Per-asset and contractor keys with expiry | Shared field credentials | Security engineering |
| Retention | Configurable logging of requests and responses | No defined schedule | Records management |
| Audit trail | Request, model and source logging | Safety decisions not traceable | HSE |
| Grounding | Embeddings and RAG over procedures and equipment docs | Outdated field guidance | Engineering |
| Review | Citations and structured output for engineers | AI guidance used unreviewed | Asset manager |
Frequently asked questions
Does using Plugsky make us compliant?
No. Compliance is your program. Plugsky provides deployable controls - scoped keys, private placement, logging - that you document and audit against your own safety, JV and regulatory obligations.
Can data stay at the field site?
Yes. On-prem, edge and air-gapped deployment keeps prompts, documents and embeddings inside the site or operating company boundary, and the API stays OpenAI-compatible.
What should we log?
Request IDs, model names and versions, retrieved sources, outputs and the reviewing engineer, retained under your policy so safety and operational decisions can be reconstructed.
How do we manage contractor access?
Issue separate scoped keys per contractor with expiry dates tied to the contract, and revoke on demobilisation without affecting asset teams.
Can we use AI for equipment inspections?
Vision and file endpoints are coming soon. Today, use retrieval over inspection standards and maintenance history, with qualified staff retaining disposition decisions.
How do we handle JV data?
Treat partner material as its own class, keep it in separate indexes, and document which partner may access which dataset before any workflow queries it.
Where should a pilot start?
Pilot on public standards and internal procedures with the free plan, validate answers, then move sensitive subsurface and JV content to private or air-gapped deployment.