Ollama interactive menu and slash command tutorial

This tutorial explains Ollama's interactive terminal menu and when the speaker prefers it over a direct command. Running ollama without arguments opens a launcher for chat, coding tools and agents. The right arrow opens model choices before launch. The speaker's default is GLM 5.2 cloud, so the walkthrough includes cloud choices as well as local models.

Inside a session, /model switches models, while /new starts a clean conversation when the task changes. /compact summarizes older context for longer work. The speaker cautions that this summary does not preserve every detail and should not serve as a backup. /think and /tools control thinking mode and tool use, with behavior depending on the model and launch path.

The walkthrough also covers /system for persistent working rules and /skills for listing or importing reusable instructions. /prompt displays the full prompt, tools and messages; /save writes the request JSON to a file.

The speaker reports one launcher limitation: frequent choices do not move to the top based on use history. For local AI workflows, the menu helps with discovery, but users still need to choose an appropriate model and agent. The speaker recommends keeping ollama run for scripts, automation and tasks where the exact model is already known.