Five services, offered to members and to operators evaluating membership. Every engagement
ends with you owning the result — the hardware, the configuration, the credentials, and the
documentation to run it without us.
Sovereign AI deployment
We stand up open-weight language models on hardware you control — a workstation with a
capable GPU, a colocated server, or capacity on co-op shared infrastructure. Inference
stays inside your network boundary, so prompts, documents, and customer records are
never sent to a third-party API.
It includes model selection matched to your workload, a private inference endpoint your
applications can call, and retrieval over your own document store when the model needs
grounding in your material. You get the runbook for updating, pinning, or rolling back
a model version on your own schedule.
Sovereign Stack build-out
The full self-hosted application stack, provisioned as one coherent system: compute and
container orchestration, storage and backup, identity and single sign-on, email and
team communications, plus the automation layer that ties them together.
The point is to own the whole path rather than rent each segment. We favour boring,
well-supported open-source components, document how each piece is configured, and hand
over infrastructure definitions so the stack can be rebuilt if a machine is ever lost.
Sovereign Data and migration off SaaS
Getting your data out of proprietary platforms and into formats and stores you control.
We inventory what lives where, export it, convert it into open formats, and load it into
systems you hold the keys to — with the relationships between records preserved, not
flattened into a folder of spreadsheets.
Migrations are staged so the business keeps running: the old system stays available
while the new one is validated in parallel, and cutover happens only once you have
confirmed the data is complete and correct.
Co-op advisory and architecture review
A structured assessment of your current stack: what you run, what you rent, where your
data physically sits, which dependencies would hurt most if a vendor changed terms
tomorrow, and what a realistic exit looks like for each one.
You receive a written review and a sequenced plan toward sovereignty, ordered by risk
reduction and effort, so you can start with the change that removes the most exposure.
It is the right first step if you are unsure how much of your stack you want in-house.
Shared infrastructure and member support
Ongoing membership benefits: pooled compute and storage for members who do not want to
buy their own hardware outright, shared deployment tooling and configuration templates,
and a members' channel where operators help each other work through real problems.
Members also get the co-op's accumulated operational material — hardening checklists,
backup and restore procedures, and upgrade notes — so the same problem is not solved
from scratch by every member in turn.