Local development with the kagent CLI

In this free, on-demand technical lab, learn to build, run, and deploy AI agents using the kagent CLI, including integration with MCP servers.

• Create a new agent: Initialize a Python-based agent with gpt-4, explore built-in tools like roll_die and check_prime, and interact with it locally via the terminal UI.
• Add an MCP server: Integrate the add tool from the Everything MCP server, update your agent configuration, and test tool interactions.
• Filter and manage tools: Configure which MCP tools are available to the agent and rebuild the project to apply changes.
• Deploy to Kubernetes: Use kagent deploy to push the agent and MCP servers to a local Kubernetes cluster, with secrets for API keys and proper resource configuration.
• Test the deployed agent: Interact with the agent through the kagent dashboard, confirming tool functionality and verifying results from both the agent and MCP servers.

Gain hands-on experience building AI agents, connecting them to MCP servers, and deploying them in Kubernetes using a fully integrated development workflow.

‍