In TryDirect and Stacker language, a stack is a group of services that work together to deliver one useful system.
Instead of thinking about one container in isolation, a stack treats the app, database, proxy, background workers, and supporting tools as one operating unit.
A stack is not just software that starts once. It is the whole set of pieces that have to run together, be configured together, and be supported together.
What can belong to a stack?
- a primary app such as OpenClaw or n8n
- a database such as PostgreSQL or Redis
- a vector store such as Qdrant
- a reverse proxy and domain configuration
- environment variables, secrets, and monitoring pieces
Why this matters
Teams get into trouble when they treat a multi-service system like one isolated app. Problems in one service can affect the whole workflow, which is why stack-based thinking makes debugging, deployment, and reuse much clearer.
How stacks show up in TryDirect
- in Stack Builder, where users shape the set of services and settings
- in Stacker, where the stack definition becomes explicit in configuration and CLI workflows
- in the marketplace, where a tested stack can become a reusable asset for other teams
A simple example
An internal AI workspace might use n8n, PostgreSQL, Qdrant, a chat UI, and a model gateway. None of those pieces is the whole product alone. Together, they form the stack.