What's next

You installed the CLI, created a project, chatted with an agent, and ran a workflow, all on your own machine. Here's where to go when you want to make it do something real.

You've got the basics

At this point you can:

• create a project with stackbone init,
• run it locally with stackbone dev,
• chat with an agent in the Playground,
• add and run a workflow.

Everything from here builds on those four moves.

Make your agent do more

• Agents: how an agent really works, from sessions and turns to giving it tools that run your own code.
• The SDK: the toolbox your agent code uses. Start with AI models and a database.
• Connections: let your agent reach Slack, GitHub, Gmail, and more, without handling passwords yourself.

Go further with workflows

• Workflows: steps, retries, and long waits in depth.
• Human-in-the-loop: pause a workflow until a person approves, then carry on.

Ship it

When it works on your laptop, put it in the cloud:

• The full CLI walkthrough: the same loop with more detail, including stackbone publish to go live.
• Command reference: every command, flag, and output.

Let a coding agent help

• Coding agents: install the Stackbone skills so tools like Claude Code and Cursor can write your agent with you.

Stuck on anything? Every command takes --help, and the troubleshooting page covers the common errors.