Skip to content

Details

In this class, we’ll explore Pi, a lightweight terminal-based coding agent designed to stay small at the core while becoming powerful through extensions, skills, prompt templates, themes, packages, custom models, and custom providers. Pi gives developers a hands-on way to bring agentic coding directly into their project directory, with practical controls for tools, sessions, model switching, project instructions, and automation.

Students will learn how to install and run Pi, authenticate with subscription providers or API keys, start useful coding sessions, reference files, run shell commands, manage saved sessions, and use non-interactive prompt mode for repeatable workflows. From there, we’ll look at how Pi can be customized: adding project instructions with `AGENTS.md`, using skills and prompt templates, extending Pi with TypeScript modules, and understanding when to use Pi as an interactive assistant versus a programmable agent runtime.

By the end, students will understand Pi as more than another coding chatbot: they’ll see it as a compact agent harness for building practical developer workflows, experimenting with model/provider setups, and creating reusable AI-assisted engineering patterns from the terminal.

Related topics

Artificial Intelligence
Artificial Intelligence Applications
Low-Code

You may also like