Deploy where your data lives.
DataStrict is software you run - not a service you send your data to.
Regulated teams can't route their most sensitive prompts and records through someone else's cloud. So DataStrict deploys inside your own environment - on-premises, in your private cloud, or fully air-gapped - delivered as a single container you own and operate.
Why self-hosted
An enforcement boundary can't live in someone else's cloud.
DataStrict sees everything. Every prompt, retrieval, tool call, and model output passes through the control layer before it's allowed - the most sensitive traffic in your company, concentrated at a single point.
A governance layer that runs in a vendor's multi-tenantcloud isn't a boundary - it's an exfiltration path with a dashboard. An enforcement point only belongs inside the same trust boundary as the data it protects.
So we sell you the software, not a seat in our cloud. You deploy it where your data already lives - on-premises, in your private cloud, or air-gapped - behind your own firewall, under keys held in your own KMS. We never receive a byte of governed traffic.
See what runs inside the boundary on the control layer page.
Your trust boundary
Nothing crosses this line. DataStrict never sends governed traffic out of your network.
Deployment
Runs where you need it to run.
On-premises
Deploy into your own datacenter - bare metal or your private virtualization stack. DataStrict runs as a standard container, with no special hardware and no dependency on a vendor's cloud.
Private cloud (BYOC)
Run it inside your own AWS, Azure, or GCP account - your VPC, your subnets, your IAM. We deliver the image; you own the environment it runs in and the network it can reach.
Air-gapped
No internet required. DataStrict runs fully offline - no phone-home, no license server calling out, no telemetry leaving the network. Built for classified, regulated, and isolated environments.
Single-tenant
One deployment, one customer. No shared compute, no shared database, no noisy-neighbor blast radius. Your instance is yours alone - sized, patched, and upgraded on your schedule.
Packaging
One container. Standard, boring infrastructure.
DataStrict ships as a signed OCI image. Run it the way you already run everything else: Docker on a single host, a Helm chart on your Kubernetes cluster, or your own orchestration.
No bespoke installer, no agent sprawl, no kernel modules. If your platform team can run a container, they can run DataStrict - and scan it, pin it, and roll it back like any other image in your supply chain.
Walk through a self-hosted install in the documentation.
# Docker - one host, behind your firewall
docker run -p 8443:8443 \
-v ./policies:/etc/datastrict \
datastrict/gateway:latest
# Kubernetes - your cluster, your namespace
helm install datastrict datastrict/gateway \
--set ledger.database="your-postgres"
# Air-gapped - load from a tarball, no registry
docker load -i datastrict-gateway.tarData sovereignty
You own the deployment - and everything in it.
You hold the keys
Encryption keys, model credentials, and TLS certs live in your KMS and never reach us. DataStrict can't decrypt what it never holds.
The Ledger stays in your database
The append-only, hash-chained record of every decision is written to a database you run. Your audit evidence never leaves your jurisdiction.
No phone-home
DataStrict needs no outbound connection to operate. Updates are images you pull and verify on your own schedule, not pushes we make on ours.
Govern AI like infrastructure.
Talk to our team about deploying DataStrict across your enterprise stack.