Real-time token estimation as you type. See how much context window you're using and what it costs per request.