tokenstat

You run several AI coding agents across several machines. tokenstat puts the whole picture in one place: token usage, what each subscription is really worth, and live control of the work while it runs. Monitor - A year-shaped heatmap of your AI usage across every machine and harness, with today and this week at a glance. - Breakdowns by model, tool, project, and day, so you can see which agent eats most of your allowance. - See your plan usage at list rates, so you can compare subscription value against API pricing. - Live plan limits for the tools you run, with when each window resets. Reach - Every machine in one list, awake or asleep, reachable from anywhere. - Start a session on any of eighteen agent harnesses. - Chat with a workspace agent in Plan, Execute, or Ask mode, and switch models mid-thread. - Review diffs, commit, and read pull requests with their conversation and checks.…