Skip to main content
Once you have set up the prerequisite agent CLIs and installed TermMate, you are ready to start chatting.

​
1. Start Chat

Once TermMate is installed, open a chat session from within your editor.

​
Sublime Text

  • Open the Command Palette (Cmd+Shift+P on macOS, Ctrl+Shift+P on Windows/Linux).
  • Type TermMate: Start Chat and press Enter.
  • A new view will open for the TermMate chat.
  • Type your message and press Cmd+Enter (macOS) or Ctrl+Enter (Windows/Linux) to send. You can even edit prompts in Vim mode.

​
VS Code

You can start a chat session in two ways:
  • Status Bar: Click the TermMate item in the bottom status bar.
  • Command Palette: Open the Command Palette (Cmd+Shift+P on macOS, Ctrl+Shift+P on Windows/Linux) and run TermMate: Start Chat.

​
2. Select Agent

When initiating a chat session, TermMate presents a native selector interface to choose your preferred underlying AI agent:
  • Sublime Text: You can switch your active agent by clicking the Agent button located directly above the chat input area inside the chat view tab. This opens a selection list panel to choose your desired agent.
  • VS Code: You can select or switch your active agent directly via the dropdown menu located at the bottom of the chat input area inside the Webview tab. Additionally, the bottom editor Status Bar displays the current active agent and model (e.g. ▣ claude(default)).

​
Supported Agents

  • Claude Code: Powered by Anthropic’s terminal agent. Highly recommended for advanced reasoning, complex multi-file generation, and deep workspace context understanding.
  • Codex: Powered by OpenAI’s coding companion. Excellent for precise low-latency edits and interactive codebase exploration.
  • OpenCode: Powered by Anomaly. Open-source agent supporting multi-model backends, interactive code editing, and server-managed session state.
  • Pi Agent: Powered by Earendil (pi.dev). A lightweight, extensible coding agent supporting multiple model providers.
TermMate automatically sets the working directory to your current project root. The agent can read, edit, and run commands within that scope. See Directory & Projects for details on switching folders and multi-root workspaces.