Key facts
| API compatibility | OpenAI-compatible /v1/chat/completions; drop-in base URL change |
| Models | 30+ models behind one API; open-weight options for private deployment |
| Typical patterns | Unclassified doc search, logistics summaries, training drafts |
| Integration path | Connects to document and ERP systems inside the controlled environment |
| Pricing model | Flat monthly self-serve plans with unlimited fair-use usage; no per-token billing — see the live pricing page |
| Free tier | plugsky-micro and plugsky-lite on the free plan, no card required |
| Deployment | Plugsky cloud, your VPC, on-prem and air-gapped options |
| Compliance posture | SOC 2 Type II and ISO 27001 readiness in progress (not yet certified); validate evidence during diligence |
TL;DR
- Limit scope to unclassified, administrative workloads.
- Deploy on-prem, air-gapped or sovereign; no external calls.
- Enforce classification-aware access and full accountability.
- Keep command and decision systems outside AI scope.
- Start free with plugsky-micro and plugsky-lite, no card required.
How it works, step by step
- Confirm with your security authority which data classes and networks are in scope.
- Choose an on-prem, air-gapped or sovereign deployment that meets accreditation.
- Define role-based access that matches clearances and need-to-know.
- Build against the OpenAI-compatible endpoint inside the boundary.
- Agree an offline update and support process before go-live.
- Pilot on unclassified administrative work with named reviewers.
Original data
Try it yourself
Open the private LLM cost estimator →
Where an AI API fits in defense
Where connectivity is restricted, the practical value is in reading and writing at scale. Suitable unclassified starting points:
- Unclassified document search: retrieve and cite policy, technical and administrative documents inside the enclave.
- Logistics summaries: condense supply and maintenance reporting into reviewable summaries for planners.
- Training material drafts: turn approved source content into draft courseware for instructors to edit.
- Procurement correspondence: summarise long solicitation threads into action lists with owners.
- Translation support: draft translations of unclassified administrative text for human review.
Security, privacy and data handling
The controlling requirement is the boundary, not the model:
- Keep all processing inside the accredited environment; no external calls from sensitive networks.
- Enforce classification-aware access so users only reach content they are cleared for.
- Apply export-control and supply-chain review to every model and dependency you deploy.
- Log requests and outputs for accountability, and keep humans approving every result.
Deployment options and model choice
The accredited boundary decides the tier: on-prem or air-gapped inside the enclave, with no external calls from sensitive networks. The same OpenAI-compatible API runs across Plugsky cloud, a private endpoint in your VPC, on-prem and air-gapped, with region-locked planes for residency. One key reaches 30+ models, including open-weight options for offline deployment, and migration is a base URL change. Chat, streaming, JSON mode, function calling, embeddings, RAG and agents are live; audio, images, moderation, files, batch, fine-tuning, assistants and the responses API remain coming soon. The free plan includes plugsky-micro and plugsky-lite with no card, and a 14-day full-access trial covers paid tiers — see the live pricing page for current plans.
From pilot to production
In restricted environments, the failure mode is often an unauthorised convenience path:
- Letting a pilot quietly use an internet-connected endpoint for convenience.
- Assuming access control maps to clearances without verifying it.
- Skipping export-control review of model provenance and dependencies.
- Treating translated text as authoritative without linguist review.
- Deploying without a support and update process that works offline.
Define the enclave boundary, access model and offline update process before any pilot. Keep the first workloads unclassified and administrative, with named reviewers, and expand only after the security authority signs off.
Honest comparison
| Capability | Plugsky | Typical per-token API | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible chat, embeddings and tools | Usually compatible | Full rewrite |
| Deployment | Cloud, VPC, on-prem and air-gapped | Mostly cloud-only | You operate GPUs and serving |
| Data residency | Region selection and sovereign options | Limited regions | You control fully |
| Pricing | Flat monthly self-serve, fair-use usage | Per-token, harder to forecast | GPU plus operations cost |
| Model choice | 30+ models behind one API | Varies by provider | You host every model |
| Industry fit | Unclassified doc search, logistics summaries, training drafts | Generic API, you adapt it | You build every workflow |
Frequently asked questions
Can we keep our existing OpenAI SDK code?
Yes. Plugsky exposes an OpenAI-compatible API, so the same application code works across deployment tiers, including inside a controlled environment.
Is there a free plan?
Yes. The free plan includes two free models, plugsky-micro and plugsky-lite, and does not require a credit card.
Can it run fully air-gapped?
Yes. Air-gapped deployment is available for enterprise setups where no traffic may leave the environment; on-prem is available where connectivity is restricted.
How does pricing work?
Self-serve plans are flat monthly with unlimited fair-use usage; enterprise agreements cover private, sovereign and air-gapped deployment. See the live pricing page for current plans.
Which workloads should stay out of scope?
Command, targeting, weapons and other decision systems. Keep the API to documents, logistics, training and correspondence with human review.
Which endpoints are live today?
Chat completions, streaming, JSON mode, function calling, embeddings, RAG and agents are live. Audio, images, moderation, files, batch, fine-tuning, assistants and the responses API are coming soon.
Is Plugsky certified for defense workloads?
No certification claim should be assumed. SOC 2 Type II and ISO 27001 readiness are in progress; validate controls and accreditation requirements directly with the enterprise team.