Your first conversation and picking a model

What you’ll learn

This is the first hands-on lesson of the Getting Started track and the one the rest of the track stands on. You will open Clawless, send your first message, and learn the model picker, the dropdown chip in the dock row that lets you choose which AI model handles each message. You will see how to switch models in the middle of a conversation without losing your place, why there is no per-conversation lock, and how to reach a model that is not on the curated short list. By the end you will have a working mental map of the four screen zones (agent rail on the left, the chat in the middle, the dock row below the input, and the navigation rail at the far left) and a feel for how the cost meter, the header label, and per-agent preferences actually behave when you click around.

Before you start

Prerequisites: lesson 1, AI won’t replace you, but it will expose you, where you picked one recurring task and split it into mechanical and judgment parts. You will use that task as the conversation starter for the practice. You also need Clawless installed and the four-step onboarding finished, with at least one provider connected (an API key or a ChatGPT OAuth session). If onboarding is not done, finish it first and come back; the lesson assumes the chat surface is in front of you.

By the end, you’ll be able to

• Identify the four screen zones of Clawless (agent rail, chat, dock row, navigation rail) and the role each plays
• Send a first message and use the activity indicator and Stop button to control the response stream
• Switch models mid-conversation using the model picker, and explain why there is no per-conversation model lock
• Use the provider-prefixed pattern (provider slug, slash, model name) to select a model that is not on the curated short list
• Predict how the cost meter, the chat header label, and per-agent preferences behave when you switch models or switch agents

Time and difficulty

• Read time: about 11 minutes
• Practice time: about 20 minutes (a hands-on with three real messages, one mid-conversation model switch, and one off-list model, ending with the Usage dashboard for the real cost numbers)
• Difficulty: intro