Self-hosted

Run Stategraph in your own infrastructure.

Docker Compose, Kubernetes, Amazon ECS, or Google Cloud Run. The same CLI, API, and features as the hosted service — with your state and data entirely inside your own environment.

Docker Compose Kubernetes Amazon ECS Google Cloud Run Air-Gapped
Get the server image Deployment docs

Where Stategraph runs

Run it managed, run it yourself, or have us run it inside your account.

Managed

Stategraph Cloud

The fastest way to start — we run the server, upgrades, and availability. See pricing →

We run it, in your cloud

Bring Your Own Cloud

We operate Stategraph inside your AWS, GCP, or Azure account. Enterprise →

Four ways to self-host

Pick the method that fits your stack. The server is a container backed by PostgreSQL — it runs anywhere you run containers.

Docker Compose Development · small teams · evaluation

The quickest path to a running server. One compose file brings up the server and its PostgreSQL database for evaluation or a small team.

Docker Compose guide →
Kubernetes Production · scaling · existing K8s

Deploy with the official Helm chart. Native liveness and readiness probes plug straight into your cluster's health checks and rollouts.

Kubernetes guide →
Amazon ECS AWS-native · managed containers

Stand it up on ECS Fargate with the official Terraform module — managed container orchestration without running a cluster yourself.

Amazon ECS guide →
Google Cloud Run GCP-native · serverless

Run the server as a serverless, fully managed container on Cloud Run — scale to your load with no infrastructure to manage.

Cloud Run guide →
The server image, Helm chart, and ECS Terraform module are distributed privately, so we can make sure your team has the support and onboarding it needs. Request access →

The same Stategraph, in your environment

Self-hosting changes where it runs, not what you get.

Full feature parity

CLI API Console
identical to hosted

The identical CLI, API, and console. Velocity, Insights, Inventory, Cost, Orchestration, and AI Agents all work exactly as they do on Stategraph Cloud.

Your data, your perimeter

Your infrastructure
state · database · never leaves

State lives in a PostgreSQL database you run. Nothing phones home; air-gap it entirely if your compliance posture requires it.

Credentials stay local

CLI Your Cloud
server never sees them

The CLI runs Terraform/OpenTofu locally against your cloud with your own credentials. The server coordinates state — it never runs Terraform and never sees your cloud credentials. More on security →

Built to operate

The hooks your orchestrator already expects.

stategraph — operations
[liveness] GET /health/live — for orchestrator restart probes [readiness] GET /health/ready — 200 only when the backend is serving [config] environment variables — no config files to template [state] PostgreSQL — your database, your backups [auth] Local, Google OAuth, or OIDC — wire in your IdP [upgrades] pull a new image tag on your own schedule

Run it your way

Tell us your stack and compliance posture and we'll send the server image, the Helm chart, and the ECS module — and get you running.

Get the server image Deployment docs