Orch

Environment orchestration

Bring up real environments. Tear them down cleanly.

Orch orchestrates component execution across local and remote runners, capturing state as it goes so every environment can be inspected, retried, and torn down reliably.

Get startedRead concepts

Terminal

$ orch up —env-id pr-123
[local > setup] apply started
[local > setup] outputs captured

$ orch state inspect —env-id pr-123
component  status   stage
setup      applied  post_apply

$ orch down —env-id pr-123
[local > setup] destroyed
state bundle deleted

One manifest for the messy middle.

Real environments rarely belong to one tool. Orch gives those pieces a shared lifecycle so setup scripts, containers, infrastructure modules, cloud stacks, hooks, and outputs can move together.

sh

Scripts

Run setup, glue code, checks, migrations, and custom automation.

DC

Docker Compose

Bring app services up locally or on a remote runner.

TF

Terraform

Apply modules, capture outputs, and preserve local state artifacts.

CF

CloudFormation

Create AWS stacks and feed stack outputs to later components.

orch.yaml

components:
  web:
    type: docker-compose
    runner: local
    source:
      files:
        docker-compose.yml: ./compose.yml

  smoke:
    type: script
    depends_on: [web]
    env:
      BASE_URL: "http://localhost:${web.outputs._meta.ports.services.web.80}"
    source:
      embedded: |
        curl -fsS “$BASE_URL/health”

Readable enough to review. Structured enough to recover.

The manifest stays close to the way developers already describe environments: runners, components, source, outputs, hooks, and state backend configuration.

Components can depend on outputs from earlier components.

Adapters capture payloads and artifacts needed for teardown.

State backends keep runtime facts outside the manifest.

Designed around lifecycle, not just launch.

Orch tracks where each component is in the run, captures the operational facts needed to destroy it, and keeps enough context to retry after interruption.

01

Declare

Describe components, runners, inputs, outputs, and hooks.

02

Apply

Run components in dependency order across your runners.

03

Capture

Persist non-sensitive outputs, payloads, and tool artifacts.

04

Inspect

See component status, stage, runner, and recovery hints.

05

Destroy

Restore artifacts and tear down in reverse state order.

Explore the docs

QuickstartRun a local script component, inspect state, and tear it down.State and recoveryUnderstand statuses, stages, artifacts, and retry behavior.AdaptersUse Script, Docker Compose, Terraform, and CloudFormation.SecurityLearn how Orch treats state, outputs, SSH, and credentials.