About this project
TokenBar is a native macOS menu bar application that monitors AI token usage and subscription quotas for over 25 AI coding agents, including Claude Code, Codex, Cursor, OpenCode, and Gemini CLI. It reads data locally from session logs, requiring no account or telemetry. The menu bar displays today's tokens, cost, live tokens per minute, or remaining subscription quota, with visual indicators like signal bars, a ring, or a melting popsicle. A cat animation runs faster as usage increases, inspired by RunCat.
The dashboard opens as a Liquid Glass panel (popover on older macOS) with app tabs to filter agents and a view switch offering seven lenses: Overview, Models, Monthly, Daily, Hourly, Stats, and Agents. It also includes an interactive 3D contribution graph for yearly usage, OAuth quota cards with pace projections, live session traces, streaks, and full keyboard control. Failed refreshes retain the last known value.
Installation is via Homebrew cask: `brew install --cask nanako0129/tokenbar/tokenbar`. The app requires Apple Silicon and macOS 14+, with Liquid Glass effects on macOS 26 and the glass panel on macOS 27. It is ad-hoc signed, and updates are delivered via Sparkle with an optional beta channel. A legacy Tauri build is available for macOS 11–13.
Architecture: Rust handles data parsing and aggregation via the shared tokscale-core engine (pinned as a Git submodule), with an FFI layer for quota fetching. Swift manages the UI, NSStatusItem shell, and Sparkle updates. Build with `make` or `make run`. The project includes a knowledge base and contribution guidelines.
TokenBar is MIT-licensed and builds on tokscale by Junho Yeo, with roots in tokcat and RunCat. Sponsorship is available via Patreon.
Comments
0 Rating appears after 10 ratings
Sign in to join the discussion.