Tactical Intelligence
Workstation
Air-Gapped Local LLM Infrastructure for Defence Operations
A secure, air-gapped-ready intelligence workstation enabling on-device processing of tactical intelligence, doctrine-aware reasoning, and visual analysis. All inference executes locally on operator hardware — no external API calls, no telemetry, no cloud dependency — with complete data sovereignty.
Tactical RAG
Ingest mission SITREPs, fragmentary orders, and geospatial overlays for instant retrieval and contextual cross-analysis against the live mission archive.
Doctrine-Aware Reasoning
Automated cross-referencing of tactical data assets against MOD Joint Service Publications and active Rules of Engagement to flag compliance deviations.
Visual Intelligence
Fully local analysis of inbound tactical imagery, structural landmarks, and potential threat corridors via on-device vision models with zero data exfiltration.
Secure Reporting
One-command generation of formal SITREPs as sanitized PDF readouts, with automated 60-second secure file purge following document creation.
To update the persistent doctrine knowledge base with JSPs or ROE cards, use the /policy route token prepended to your query:
Query both archives concurrently using natural language. The system forms validated cross-references across mission data and doctrine automatically.
Execute the Generate Formal SITREP route to produce a sanitized, standalone PDF readout from the current session context.
OPSEC Protocols
This node binds exclusively to loopback interface 127.0.0.1. External tracking telemetry routes are structurally omitted — not disabled, but absent.
All attached source documents are securely shredded via a 3-pass sector-level wipe following embedding completion. No residual data persists on disk.
Heavy models (Gemma 27B, Moondream) execute natively within onboard GPU frameworks. No tokens leave the host machine at any point.
Hardware Requirements
| Parameter | Minimum | Recommended |
|---|---|---|
| VRAM / RAM | 16 GB | 24 GB+ |
| Storage | 20 GB free | 40 GB+ SSD |
| GPU (NVIDIA) | RTX 3060 16GB | RTX 4080+ |
| Apple Silicon | M1 24GB | M3 Pro/Max 48GB |
| Quantization | IQ4 / 4-bit | Q6_K / 8-bit |
Select your target architecture below. Each request routes a deployment bundle inquiry directly to the system administrator and includes pre-configured Ollama manifests, model download scripts, and setup documentation.