mm7894215/TokenTracker

Local-first AI token usage & cost tracker for 27 coding tools — with a desktop pet, 4 widgets, achievements, native macOS/Windows apps, and a one-command CLI. Never reads prompts.

What it solves

TokenTracker provides a centralized, local-first dashboard to track token usage and costs across dozens of different AI coding tools. It eliminates the need to check multiple provider dashboards and provides a unified view of spending and rate limits without requiring cloud accounts or API keys.

How it works

The tool uses a combination of lightweight hooks, plugins, and passive readers to monitor AI CLI tools. It parses local logs, SQLite databases, and session files produced by these tools to extract token counts and timestamps. This data is aggregated into 30-minute UTC buckets and stored in a local SQLite database, which then feeds a web dashboard, a menu bar app, and desktop widgets.

Who it’s for

Developers who use multiple AI coding assistants (such as Cursor, Claude Code, GitHub Copilot, and others) and want to monitor their token consumption, costs, and rate limits in one place.

Highlights

  • Broad Support: Supports 27 AI tools out of the box, including major providers like Claude, Gemini, and GitHub Copilot.
  • Local-First Privacy: Token data stays on the machine; prompts and responses are never uploaded.
  • Zero Configuration: Hooks are automatically installed on the first run.
  • Cost Engine: Includes pricing for over 2,200 models via LiteLLM with daily auto-refreshes.
  • Gamification: Features a desktop pet that reacts to coding activity, 15 achievement tracks, and an optional global leaderboard.
  • Native Integration: Offers macOS menu bar and Windows system tray apps with native desktop widgets.