Skip to content

Getting Started

Launching

Launch from your application menu or run:

ctfl

CTFL starts as a system tray icon.

Basic interactions

  • Left-click the tray icon to toggle the usage popup
  • Right-click for the context menu (refresh, settings, check for updates, quit)
  • Hover over the icon to see a quick summary in the tooltip

Tray tooltip

Usage popup

The popup shows your token usage in several views:

  • Daily — bar chart of tokens per day
  • Models — breakdown by Claude model
  • Projects — breakdown by working directory

Daily usage Per-model usage

Long lists scroll inside the tab instead of stretching the popup off-screen, so raising Days to show in Settings is safe.

Above the charts, the popup surfaces plan rate-limit bars fetched from claude.ai. Pro and Max plans show session (5-hour) and weekly bars; Enterprise plans show a monthly spend bar with used / cap credits. The same figures appear in the tooltip when it is enabled.

Long-context usage hint

When recent activity is dominated by large conversations, the popup surfaces a hint under the period total, for example:

Recent sessions: 61% of tokens used at >150k context · /compact mid-task, /clear between tasks

Each message you send to Claude re-includes the whole conversation. Once the context passes ~150k tokens, each further reply is much more expensive — even with prompt caching. If you see this hint, running /compact mid-task or /clear when switching to unrelated work significantly reduces rate-limit burn.

The hint is computed over recent (JSONL-logged) sessions, not the full displayed period — older data aggregated by Claude Code's stats cache doesn't include per-message context size, so only the recent window can be measured.

Next steps