What We Offer
Managed Private AI Nodes
We supply, install, and maintain enterprise-grade AI hardware in your office. You get the power of modern AI — with zero cloud dependency.
Monthly Plan
Managed Maintenance
A single, predictable monthly fee covers everything needed to keep your AI node running at peak performance — so your team can focus on client work, not IT.
Monthly maintenance from
£199/month
Remote Health Monitoring
24/7 monitoring of hardware temperatures, memory usage, and service uptime. We catch issues before you notice them.
Model & Firmware Updates
We push the latest open-source model releases and security patches to your node, tested and verified before deployment.
UK-Based Telephone & Email Support
Talk to a real engineer, based in the UK. No overseas call centres. Response within one business day guaranteed.
Security & Compliance Reviews
Annual review of your node's access controls and network configuration to maintain your GDPR and ICO compliance posture.
The Hardware
Enterprise AI, Sized for an SME
Powered by NVIDIA's RTX 5090 — the same GPU architecture used in research and enterprise AI deployments, in a compact workstation that fits under a desk.
32 GB
GPU VRAM
Runs 70-billion parameter models at full quality. Handles large legal documents, complex contracts, and financial reports without truncation.
100+
Tokens per Second
Responses generated at human reading speed or faster. No waiting for cloud round-trips — output appears in real time.
Air-Gap
Ready
Can operate on a fully isolated network with no internet connection. Suitable for the most sensitive compliance environments.
In plain terms: Your node can read a 200-page contract, summarise it, extract key clauses, and flag risk areas — all in under a minute, entirely on your own hardware. No document ever leaves your building.
Getting Started
From Enquiry to Live AI in Weeks
A straightforward process with no disruption to your team.
Discovery Call (30 min)
We discuss your team size, existing workflows, compliance requirements, and what you want to achieve with private AI. No technical knowledge needed.
Proposal & Agreement
We send a written proposal including hardware specification, one-time cost, and monthly maintenance fee. You review, agree, and we invoice. Simple.
Hardware Build & Dispatch
Your node is built, tested, and pre-configured with your selected AI models before dispatch. Typical build time is 5–10 working days.
On-Site Installation & Handover
Our engineer installs the node, connects it to your network, and runs a live demonstration with your team. You leave the session ready to use your private AI on day one.
Ongoing Managed Support
From month one, your maintenance plan kicks in. We monitor, update, and support your node remotely — you just use it.
AI Models
Best-in-Class Open Models, Privately Hosted
We configure and manage the models best suited to your workflows. These run entirely on your hardware — no licensing fees, no usage caps.
Llama 3.3 70B
Best for: Contract review, legal research, complex document summarisation
Meta's flagship open model. Exceptional at understanding long legal documents, identifying obligations and risk clauses, and drafting professional correspondence.
Qwen 2.5 32B
Best for: Financial data analysis, spreadsheet reasoning, report generation
Outstanding numerical reasoning and structured data understanding. Ideal for accountants reviewing financial statements, preparing management accounts, or analysing client portfolios.
Mistral Small 22B
Best for: Fast Q&A, client email drafting, quick document queries
Fast, efficient, and highly capable for day-to-day tasks. Perfect for team members who want quick answers or need to draft routine client communications at speed.
Phi-4 14B
Best for: Research tasks, policy interpretation, structured reasoning
Microsoft's compact reasoning model. Punches well above its size for research synthesis, regulatory guidance interpretation, and step-by-step analytical tasks.
Additional models available on request. We keep up with the rapidly evolving open-source AI landscape so you don't have to.
Ready to get started?
Book a 30-minute discovery call and we'll walk you through exactly how a VaultLLM node would integrate into your practice.
Book a Discovery Call