AI that never leaves
the building.
Same workflows, run where your data is allowed to be: private cloud, on-premise, or air-gapped. Built for law firms, clinics, finance, public bodies and anyone whose regulator or client contract rules out a shared platform.
This is the enterprise lane and it is sales-led on purpose. It starts with a fixed-fee assessment, not a signup form. Read Sovereign in full →
Assessed, proven, then live.
It starts with an assessment, not a signup form, and it is backtested on your own history before it touches live work.
Pick a scenario, or any box.
Scroll the diagram sideways to see the whole path.
The line item it takes off your desk.
The cloud you cannot use
The front desk, back office and portal workflows on infrastructure you control, grounded in your own documents, with citations.
The custom build nobody scoped
Six productised solutions to start from. Blueprint first, fixed price, credited to the build.
The hardware you would have to source
Edge supplies the on-prem AI box, scanners and sensors already paired to the software. One point of contact for both.
It has to earn the switch.
It reproduces your own past before it runs, a person approves what it drafts, and the log keeps every decision. The six controls →
A Sovereignty Assessment first
A written read on whether private AI fits, which tier suits your data, and what a pilot looks like.Fixed fee from $5,000, credited to the build
The same gates as the platform
Nothing is relaxed because it is private. Approvals, backtests, the safety battery and conversation scoring run identically on-prem.
Honest about certifications
We do not claim certifications we do not hold, and we will answer any questionnaire in writing.
What you actually see.
It starts with a person and an assessment.
Your data classes, what constrains them, the tier that fits.
Architecture, acceptance tests, a fixed price.
On your infrastructure, backtested on your history.
Nothing trains a shared model.
Invoiced monthly, under one master agreement.
Not self-serve
Not self-serve. There is no free tier for private deployments and no signup form; it starts with a person and an assessment.
Only the assessment fee is published
Build and monthly costs are scoped per engagement after the assessment. We publish the assessment fee and nothing else, because the honest number depends on what you already have.
Open-weight models are smaller than the largest cloud models
Open-weight models on your hardware are smaller than the largest cloud models. For most document, intake and back-office work that is fine; for some tasks it is not, and the assessment says which.
On-prem means your team owns the box
On-prem means your team owns the box. We document, monitor and support it; we do not pretend it is zero-maintenance.
No certification we do not hold
We do not claim any certification we do not hold, and we answer questionnaires in writing.
Canada and EU residency is a deployment choice
Data-residency for Canada and the EU is a deployment choice here, not a checkbox on the shared platform yet.
Assessment first. A scope you can review.
Private infrastructure is a separate engagement. The assessment defines the build, hosting and support requirements. Every tier, in your currency →
Sovereignty Assessment
Fixed fee, credited to the build. A written recommendation, the tier and the pilot scope.
Build and run
Fixed-price build from the blueprint, then a committed annual engagement invoiced monthly. Multi-site priced per site, not per seat.
Try it on your own questions.
Draft-first from the first minute. Nothing sends, pays or files until your team says so, and your data is never used to train a shared model.