# TokenBar

> A local-first AI token usage monitor for the macOS menu bar. It reads your
> AI coding session logs on-device and shows what you spend across 25+ coding
> agents — no telemetry, no account. Built with Tauri 2 (Rust + React/Vite).

## What it is
- macOS menu-bar app for Apple Silicon, macOS 11+. MIT licensed, free, open source.
- Tracks 25+ AI coding agents from local logs: Claude Code, Codex CLI, Cursor, OpenCode, Gemini CLI, Copilot CLI, Amp, Droid, Hermes, Goose, Kilo/KiloCode, Roo Code, Qwen, Kimi, Crush, Zed, Kiro, Trae, Warp, and more.
- Session parsing, dedup, and pricing are delegated to the vendored tokscale-core crate.

## Dashboard views
- Overview: contribution chart (2D stacked bars + interactive 3D year graph), agent limits, live session trace, model breakdown, streaks.
- Models: every model ranked by cost with an In/Out/CacheRead/CacheWrite split.
- Daily: active days, selectable to drill into that day's per-model breakdown.
- Hourly: a 24-hour-of-day rhythm of token spend.
- Stats: total spend, active days, streaks, favorite model, best day.
- Agents: sub-agents ranked by cost with source apps, message count, and tokens.

## Privacy
- Usage history is read locally from disk. No telemetry, no analytics, no cloud sync, no account.
- Network access is limited to the GitHub Releases updater manifest and public model-pricing data.

## Install
- Homebrew: `brew install --cask nanako0129/tokenbar/tokenbar`

## Links
- Repository: https://github.com/Nanako0129/TokenBar
- Forked from tokcat by handlecusion: https://github.com/handlecusion/tokcat
- Parsing powered by tokscale by Junho Yeo: https://github.com/junhoyeo/tokscale
