Want to know how to run AI tools directly from the command line instead of the browser? In this article, we cover OS support, WSL2 setup, and installing Gemini command-line interface (CLI), ChatGPT Codex, and Amazon Q (now Kiro CLI). With terminal access, AI can work with files, generate and refactor code, and run toolchains while maintaining project context. In one example, Gemini CLI built an Arduino library end-to-end.

AI Inside

Most people use AI entirely inside a browser; and for short questions, this works fine. The problem starts when you are deep into a research or engineering task. For AI-supported programming work, this usually means uploading all the relevant material, reading documents, extracting notes, and then copying the AI’s output back into your own files and tools. Code snippets are pasted into the IDE, compilation errors appear, and the error output is then copied back into the chat to ask for fixes. The response is applied manually, and the cycle repeats. As t...