You own the model
The AI runs on infrastructure you control. When you stop paying us, it keeps running. No subscription kills the intelligence you built on your own data.
Governed AI
Own your AI.
Your data.
Your infrastructure.
We deploy private, governed AI on hardware you control. Your sensitive data never leaves your environment — we never route it to OpenAI or Google. Yours from day one.
What is private AI, and why does your business need to own it?
Private AI means your models run on your hardware — prompts and documents never leave your network. Governed AI wraps every decision in a tamper-evident audit trail: logged, attributable, bounded by rules you define. Own the intelligence of a frontier model without surrendering sensitive data to someone else's cloud.
What does it mean to own your AI?
Four guarantees that distinguish a governed deployment from a SaaS subscription or a shared-cloud API.
Your data never leaves
Inference happens on your hardware. Prompts, documents, and responses never travel to OpenAI, Google, Anthropic, or any other cloud provider.
Every call is logged
ORA and PulZ wrap every AI action in a tamper-evident audit chain. You can prove what the model was told and what it returned — in a compliance review or in court.
Stays inside your perimeter
GDPR, HIPAA, Bill C-27, and the EU AI Act increasingly require data to stay inside defined boundaries. A self-hosted model satisfies that requirement structurally, not by policy.
Which tier is right for you?
Three configurations. Fixed prices, published openly. No discovery call required before you know what it costs.
Self-Host Starter
$1,500 setup · then $99 / mo
Your hardware. Your rules.
We configure a private LLM stack on your own workstation or server — Ollama, model weights, the ORA governance layer, and a web UI — then hand you the keys. No cloud. No egress.
- Ollama runtime + model selection
- Browser-accessible private UI
- ORA governance layer installed
- 30-day onboarding support
- $99/mo care: updates + monitoring
VPC Deploy
$2,500 setup · then $149 / mo
Your cloud. Your VPC. Isolated.
A network-isolated deployment inside your existing AWS, Azure, or Hetzner account. No shared tenancy, no public IP. All inference traffic stays inside your perimeter.
- VPC-isolated, no public endpoint
- Private network API access
- Audit log export to your S3 / blob
- Multi-user access control
- $149/mo care: SLA monitoring + patches
On-Prem Enterprise
$3,500 setup · then $299 / mo
Airgapped. On your floor.
Full airgapped deployment on rack hardware you own, with PulZ tamper-evident audit chain, governance gates, staff training, and a signed handover checklist.
- Airgap-capable installation
- Hardware scoping consultation
- PulZ governance + audit chain
- Staff training session
- $299/mo care: on-call + SLA
All tiers include the ORA governance layer, a written quote before work begins, and infrastructure in your name — not ours. Email [email protected] to scope your configuration.
Flagship product
ORA: the governance layer inside every deployment
ORA is a security-first orchestration backend for running private agents and models with auditable gateways over HTTP, WebSocket, and MCP. It sits between your team and the model — logging every call, enforcing confidence gates, and producing a tamper-evident audit chain that holds up to compliance review.
Every private AI deployment we ship includes an ORA governance layer. It is the reason we can guarantee no decision is invisible, no prompt is unattributed, and no sensitive document can be exfiltrated without leaving a trace.
Explore ORA →From scoping call to sovereign AI in days.
Four steps. Fixed price quoted before we touch your infrastructure.
- 01
Scope the hardware
One call or email to understand your workload, user count, and compliance requirements. We recommend a hardware configuration or assess what you already have.
- 02
Fixed price, in writing
A written quote before we touch a server. Self-host, VPC, or on-prem — the price is fixed and does not move unless the scope does.
- 03
We install and configure
Ollama, model weights, ORA governance layer, and your private UI — installed, tuned, and tested against your real workload. Typically one to three days.
- 04
You get the keys
Signed handover checklist: credentials, architecture documentation, and a recorded walkthrough. Your team runs it from day one. The care plan keeps it current.
12 Apps deployed on ORA platform
$1,500 Setup — self-hosted starter
0 bytes Data sent to third-party AI
$99/mo Minimum monthly care plan
Why AI governance matters right now.
The compliance window is closing. Organizations that act early win — those that wait will scramble.
The EU AI Act entered force in 2024. Canada's Bill C-27 is advancing through Parliament. US executive orders add accountability mandates for AI in finance, healthcare, and critical infrastructure. UK regulators have signalled sector-specific rules are coming, and ICO guidance on AI data processing is already active.
Organizations running AI on their own infrastructure — with a complete, signed audit trail — can show regulators exactly what model they used, what data it touched, and what it decided. That proof is structurally impossible when you are a tenant on someone else's cloud.
EU
EU AI Act
In force 2024. High-risk AI requires audit trails and human oversight mandates.
CA
Bill C-27 (AIDA)
Canada's AI accountability framework advancing through Parliament as of 2025.
US
EO 14110
AI safety and accountability mandates for regulated sectors and federal contractors.
UK
UK AI Regulation
Sector-specific rules signalled; ICO guidance on AI data processing already active.
Questions about private AI.
What is governed AI?
Governed AI means every model decision is logged, attributable, and bounded by rules you define. ORA adds confidence gates and a tamper-evident audit trail so you can prove what the AI was told and what it decided — useful in regulated industries and anywhere accountability matters.
Will my data be sent to OpenAI or Google?
Never. Inference runs entirely on your hardware or inside your VPC. No prompt, no response, and no document you feed the model ever leaves your environment. We do not route through any third-party AI provider.
What models does private AI run?
Any model compatible with Ollama — Llama 3, Mistral, Gemma, Phi, Qwen, and dozens more. We help you select the right model for your workload: document analysis, code generation, question-answering, or internal knowledge retrieval.
Do I need a GPU server?
Depends on the model and query volume. Many businesses run 7B–13B parameter models effectively on a single mid-range consumer GPU (RTX 3090 or 4090 class). We scope the hardware requirement in the first call and tell you exactly what you need before any money changes hands.
What is ORA and why is it included in every deployment?
ORA is our flagship governed AI platform — a security-first orchestration backend that runs private agents and models with auditable gateways over HTTP, WebSocket, and MCP. It sits between your team and the model, logging every call and enforcing confidence gates. Every deployment we ship includes an ORA governance layer. See it at /software/ora.
Why does AI governance matter right now?
The EU AI Act entered force in 2024. Canada's Bill C-27 (AIDA) is advancing through Parliament. US executive orders are adding accountability mandates for AI in finance, healthcare, and critical infrastructure. UK regulators have signalled sector-specific rules. Organizations that can prove what their AI did — and what data it touched — are ahead of the compliance curve, not scrambling to catch up.
How does the care plan work?
The care plan is a monthly retainer covering model updates as new weights release, security patches, uptime monitoring, and a defined response SLA for any issue. You can cancel anytime — your deployment keeps running regardless. The base plan is $99/mo; the on-prem enterprise plan is $299/mo.
What if I already have a GPU server or cloud account?
Even better. We assess your existing hardware or cloud account in the scoping call and configure the deployment there. If it meets the requirements, the setup price stays the same — we just work with what you have.
Own the intelligence.
Start at $1,500.
One scoping call. A written quote before we touch anything. Your sensitive data stays inside your walls — guaranteed. Deployments from New Brunswick, worldwide.