Summary: Your first conversation and picking a model

The short version

Clawless puts a chat in front of you with a model picker close at hand. You type, press Enter, and a reply streams back. The picker, the dropdown chip in the dock row below the input, lets you choose a model from a curated short list grouped by provider, or type a provider-prefixed name to use an off-list model. You can switch models mid-conversation with no per-conversation lock; the new model continues where the last one left off, and the cost meter follows the model you are using at the moment you send each message. Each agent on the left rail keeps its own conversation history and its own preferred model, so clicking between agents swaps both. The header model name shows the current model you have selected. Once these four pieces (zones, picker, switching, per-agent thread) are real to you, the rest of Clawless is additions.

Core ideas

• Clawless has four screen zones: the agent rail on the left (avatars, one per agent, Assistant selected by default), the chat in the middle (header, message area, input), the dock row just below the input (New conversation button, online status dot, Skills shortcut, Tools shortcut, Commands shortcut, and the model picker), and the navigation rail on the far left (Tools, Skills, Memory, Channels, Cron, Logs, Usage, Settings panels).
• Sending a message is just typing into the input box and pressing Enter. The activity indicator below your message tells you what the agent is doing in real time. The Stop button replaces the Send button while the reply streams. Force Stop is a red banner that appears only if Stop has not taken effect after about ten seconds.
• The model picker is the dropdown chip in the dock row, displaying the current model name. Clicking it opens a curated short list of models grouped by provider, with the default selection following whichever provider key you set up first.
• The short list is intentional. Most providers ship dozens of models and the differences rarely matter for most work. The list covers a useful range: capable for hard work, fast for quick work, cheap for high-volume work.
• If the model you want is not on the short list, type the provider-prefixed name into the picker. The pattern is the provider slug, a slash, and the model name, as in:

openai/gpt-5.1-mini

• You can switch models mid-conversation. The new model picks up the full conversation history and continues where the previous one left off. There is no per-conversation model lock. This is load-bearing: it lets you use a big model for the hard first question and switch to a cheaper one for ten small follow-ups, dropping the cost per message by a factor of five or ten without restarting the thread.
• Costs follow the current model, message by message. Switching does not retroactively change anything; it changes the meter going forward.
• Each agent on the left rail keeps its own conversation thread and its own preferred model. Clicking between agents swaps both. The picker updating when you switch agents is intentional, not a glitch.
• The model name in the chat header shows the current model you have selected. If you switch models mid-conversation, the header reflects the model you have selected.
• The Usage dashboard, on the navigation rail at the far left, breaks down cost by provider, model, and conversation. Opening it after your first few conversations is the fastest way to build intuition for what each model costs you in practice.
• Slash commands (typing a forward slash at the start of the input) are a power-user shortcut, including one to switch models without leaving the keyboard. Safe to ignore for the first hour.

What changes for you

Before this lesson, choosing a model probably felt like a permanent commitment per conversation, the way “which app do I open” works in most software. Now you know the picker is a live dial, not a setup choice, and that the cost-and-capability tradeoff is something you can adjust message by message. The most useful habit forming here is starting hard, then dropping down: open the conversation on a big capable model so the first reply is strong, then switch to something smaller and cheaper once the work is mostly follow-ups. Across the rest of the track, every other Clawless feature (Skills, Tools, Memory, agent customization) sits on top of this one capability, so the time spent practicing it in the hands-on pays back across everything else.