Claude Code Usage Monitor: Live Limits in Terminal

Claude Code Usage Monitor (PyPI package claude-monitor) is a privacy-first terminal dashboard for Claude Code usage. It shows a live Rich UI with token consumption, burn rate, reset timers and forecasts of when you will hit your plan limit. Version 4 rebuilt it as a “usage-ops companion” with an official-limit trust layer, machine-readable output and an optional local history warehouse.

Who it is for

Pro and Max subscribers who regularly run into session or weekly limits and want to pace long Claude Code sessions instead of being surprised by a cutoff.

Install

The README recommends uv:

uv tool install claude-monitor
claude-monitor  # or cmonitor, ccmonitor for short

pipx install claude-monitor and pip install claude-monitor also work. Then pick your plan:

claude-monitor --plan max5 --theme dark --timezone "America/New_York"

The plan is saved between sessions. For scripts and status bars, claude-monitor --once --output json prints a single snapshot.

Use cases

• Pacing: keep the monitor in a side pane during a long refactor to see if you will make it to the reset.
• Plan sizing: watch real burn rates for a week before upgrading or downgrading.
• Automation: feed --write-state or JSON output into a tray app or tmux status line.
• History: enable the local warehouse to keep usage beyond Claude’s 30-day log cleanup and export CSV or JSON reports.

Pros

• Live, forward-looking view rather than after-the-fact reports.
• Provenance labels separate official numbers from estimates.
• Data stays local; MIT licensed.

Cons and caveats

• Estimates depend on Claude Code’s local log format, which can change between releases.
• Python tool, so it needs a Python 3.9+ environment.
• Team plan math is explicitly unverified.

Links

• Repo: Maciek-roboblog/Claude-Code-Usage-Monitor
• Related: ccusage for reports, ccstatusline for an in-session view, and our pricing guide