The infrastructure database.

Your infrastructure has a database now.

Stategraph is the operational database for infrastructure. One system of record for the resources you run, the changes you ship, and the agents making them. Queryable, governable, live.

Clients Server Capabilities API CLI HCL AI Agents STATEGRAPH Infrastructure as a database queryable · governable · live Policy Engine Remote Execution Orchestration Inventory Cost Security Compliance

You already know what's missing

The problem isn't more tools on top. Every change to your infrastructure ends up in a flat state file. Change history, drift detection, blast radius, incident review — they all inherit that limitation.

What changed?

When prod misbehaves, you need to know what infrastructure changed in the last hour. Today that requires grepping CI logs across teams.

Drift you find too late

Drift surfaces in the next plan, which might be days away. By then the gap between intent and reality has widened.

Blast radius is a guess

Knowing what depends on what is tribal knowledge. When something breaks, the answer lives in someone's head, not in a system.

Slow runs

When prod is on fire, a 15-minute plan is unacceptable. Every plan refreshes the entire state just to change three resources.

Weak visibility

You can see whether a run succeeded. You cannot query what you actually have.

Coordination cost

A cross-service fix in the middle of an incident becomes a pipeline of wrappers, waits, and rollback risk.

These are not independent problems. They are symptoms of the same missing layer: a real data model underneath.

The model is the product

Terraform is a distributed systems problem pretending to be a JSON file.

Today's tools wrap the JSON file in a UI, a runner, and a pile of add-ons. Stategraph replaces the JSON file with a real database. Every capability below falls out of that single decision.

File-based
{
  "version": 4,
  "resources": [
    { "type": "aws_vpc", ... },
    { "type": "aws_instance", ... },
    { "type": "aws_s3_bucket", ... },
    ...2,847 more
  ]
}
Read it, lock it, rewrite it. Every operation touches the whole file.
Stategraph
vpc rds eks sg net pod svc
Graph of queryable resources with real relationships. Touch only what changed.
File-based tools Stategraph
State model Flat state blob per workspace Structured, normalized database
Lock model Global workspace lock Graph-scoped, resource-level
Plan scope Refresh and plan the whole state Operate on the affected subgraph
Visibility Run history and a log viewer SQL across every resource and state
Governance Layered add-on products Capabilities that emerge from the data model
Multi-state change One workspace at a time One atomic transaction across states

One data model. Every capability flows from it.

Today's infrastructure teams stitch together half a dozen tools for plans, policy, cost, drift, inventory, audit — and now, governance over their developers' AI agents. Each one parses the same state file, badly. Stategraph collapses that sprawl into one database. Every capability below is the same data model answering a different question.

INFRASTRUCTURE as a database Agents governed by the same database Velocity graph-scoped plans Orchestration atomic cross-state apply Security blast-radius on the graph Compliance evidence by query Cost spend on real structure Inventory SQL across every state
Velocity
vpc rds sg eks
↓ plan scope: 2 of 4
plan skipped · 2

Plans and applies on the affected graph

Change three resources, touch three resources. Skip the 3,000 you did not touch. Disjoint subgraphs run in parallel without waiting on a global lock. No path globs, no change-detection config — Stategraph reads the graph and runs only what your change actually reaches.

Orchestration
PLAN ✓ REVIEW ✓ APPLY

Coordinate changes without a wrapper maze

Plan and apply across states in one atomic transaction. Dependency-aware execution, run from PRs, without hand-built pipeline glue.

Security
sg.api
↓ reaches
api-1 api-2 lb.api

Reason about real infrastructure relationships

Resource-level RBAC. Blast-radius analysis on the actual graph, not a static file scan. See what a change reaches before it ships.

Compliance
✓ encryption at rest enforced
✓ policy · restricted bucket access
▸ internal · tag coverage 98%

Query and validate state for audit and governance

Compliance and internal policy run against real infrastructure data. Full attribution, full history, exportable evidence.

Cost
aws_rds.prod$4,210 ↑ aws_eks.web$2,780 aws_s3.logs$142 ↓ Δ this week +$612

Understand cost in context of actual infrastructure

Cost attached to the resource graph, not a flat line item. See spend by change, by owner, by dependency, across every state.

Inventory
SELECT type, count(*)
FROM resources GROUP BY type;
aws_instance47
aws_s3_bucket23
aws_lambda18

Treat infrastructure as queryable data

SQL across every state, every resource, every environment. Gap analysis finds unmanaged resources and drift automatically.

Agents
human agent
↓ same policy · same audit · same approvals
infrastructure database

One database. Humans and agents share it.

Your developers' AI agents are changing infrastructure whether you authorize them or not. Stategraph is the system in between — same policy, same audit trail, same approval flow as your human engineers, applied automatically to every agent action.

None of these are add-ons. They are the data model doing its job — for every client that touches your infrastructure, human or otherwise.

Keep your Terraform. Get a database underneath it.

Adoption is incremental. Stategraph imports your existing state on day one and runs your existing HCL with a database underneath. Same providers, same workflow, same modules — now queryable, governable, and live. Move root modules over at your pace.

Terraform & OpenTofu

Same HCL, same providers, same CLI. Drop-in for the runtime you already use.

Reversible import

Ingest your existing .tfstate and export it back anytime. Never locked in.

Your HCL works as-is

Same modules, same providers, same plan and apply.

Queryable from day one

The moment state is imported, every resource is queryable. No re-tagging, no second pipeline.

Swap the backend

Point your backend block at Stategraph and ship.

migrate
$ # 1. Import your existing state
$ stategraph import terraform.tfstate
✓ Imported 847 resources. Graph built.
$ # 2. Plan on the affected subgraph
$ stategraph plan --out plan.json
✓ 3 to add, 1 to change, 0 to destroy
2 independent subgraphs detected
$ # 3. Apply in parallel
$ stategraph apply plan.json
✓ Apply complete. 4 resources updated.

Adopt at your pace. Move root modules over one at a time. stategraph states export returns any state to plain .tfstate whenever you want.

Read the migration guide Book a migration demo

Graph-scoped execution, measured

Same infrastructure. Same change. The data model does the work.

File-based backend

terraform apply 00:00
[VPC ] waiting
[Subnets ] waiting
[Security ] waiting
[RDS ] waiting
[ALB ] waiting
[ASG ] waiting
[Route53 ] waiting
[CloudFront ] waiting

Stategraph

stategraph apply 00:00
[VPC ] waiting
[Subnets ] waiting
[Security ] waiting
[RDS ] waiting
[ALB ] waiting
[ASG ] waiting
[Route53 ] waiting
[CloudFront ] waiting
Subgraph Plans scoped to what changed
Parallel Independent branches apply concurrently
Atomic Cross-state changes in one transaction
SQL Every resource queryable, every run

See how the graph model works →

Deploy where your policy requires

Stategraph is infrastructure software. It ships like infrastructure software. Enterprise control, security, and compliance teams get the deployment model they need, not the one the vendor prefers.

STATEGRAPH CLOUD YOUR INSTANCE

Stategraph Cloud

Fully managed SaaS. Free tier is shared-tenant; paid tiers run single-tenant with your own isolated instance and Postgres database. The fastest way to get started.

Start free →
YOUR NETWORK STATEGRAPH

Self-hosted

Run Stategraph in your VPC on your PostgreSQL. State never leaves your network. Full control over upgrades, secrets, and access.

Talk to sales →
YOUR CLOUD ACCOUNT WE OPERATE IT

BYOC

Bring Your Own Cloud. We operate Stategraph inside your AWS, GCP, or Azure account with private connectivity. You own the data, we run the software.

Talk to sales →

Frequently Asked Questions

See it on your infrastructure.

Bring a state file. In under 30 minutes you will see change correlation across teams, blast-radius queries against your real graph, parallel applies, and live infrastructure running against your own resources.

Book a demo Read the docs