Documentation
How the platform works, what the components are called, and how to build a workflow you can hand to an auditor. Start with the platform overview, then the quickstart.
Getting started
Platform overview
What StackHeal is, what the components are called, and how a run flows through them.
Core concepts
Assets, tenants, environments, actions, workflows, runs and policies — the vocabulary used throughout the product.
Quickstart: your first workflow
Connect a system, register an action, build a read-only workflow, and run it against one asset.
Workflows
Triggers
The six ways a workflow can start, and how payloads are bound to scope.
Actions and the action catalogue
Why workflows compose from a registered catalogue, and how to add an action safely.
Approvals and holds
Inserting a human decision into a run, with timeout behaviour that fails safe.
Verification
Why a successful action is not a verified outcome, and how to write assertions.
Retries, resume and failure policy
Being writtenRetry strategies, idempotency keys and partial-run semantics in detail.
Integrations
Connecting a system
Credentials, connectivity and the difference between a native connector and an API integration.
Generic HTTP action
Calling any authenticated endpoint as a workflow step, with typed inputs and response mapping.
Inbound and outbound webhooks
Being writtenSignature verification, payload mapping, retry semantics and event types.
Platform
Deployment models
StackHeal cloud, private execution runners and self-hosted deployment.
Security and permissions
Roles, separation of duties, secret handling and what the interpretation layer can and cannot reach.
Run history and audit
What is recorded per run, retention, and exporting for audit.
Platform status
Where to check service health, and how status notifications are published.
API
What this documentation covers
These pages describe the StackHeal platform's design: its components, controls and API shape. This website is the marketing and documentation front end — it does not execute workflows or connect to infrastructure. Pages marked “being written” are not stubs: they are listed so you can see what is missing rather than discovering it later.