Playground Workflow
A Knowledge.Dev playground runs inside your AI agent. Before you begin, make sure the local MCP Server is connected.
What It Is
This is the main way to take a full Knowledge.Dev course.
The lesson runs inside your AI agent, while the connected MCP Server provides the structure, context, and progress flow behind the session.
When to Use It
- Use it when you want the full course experience
- Use it when you want reliable lesson context and progress tracking
- Use it when your agent already has the MCP Server connected
- Use Demo Prompt if you only want a fast preview first
How to Use It
Follow the steps below and you can begin a full playground session in your local workspace.
Step 1. Create a New Folder
Create a new empty folder for the playground you are about to take.
This keeps the workspace clean and gives the agent a dedicated place to create and modify files.
Step 2. Make Sure MCP Is Connected
Before starting the playground, make sure your agent has the Knowledge.Dev MCP Server connected.
If you have not done that yet, see the MCP Server section.
Step 3. Start Your Agent
Start your AI coding agent in that folder.
This can be a CLI agent or an IDE-based agent, as long as it has the MCP Server available in the session.
Step 4. Choose a Playground
Open the playground catalog and choose the course you want to start.
Each playground has its own unique ID on its description page, and that is the value you will use in the next step.
The catalog is available at knowledge.dev/playground.
Step 5. Enter the Playground Prompt
In the agent chat, enter:
playground <id>This is not a shell command. It is the instruction prompt that tells the connected agent to open the selected playground.
Step 6. Follow the AI Tutor
After that, the AI tutor will load the playground instructions and begin guiding you through the session.
Follow the instructions, write the code, and build your application step by step.
If you have questions, visit our Discord