Tracking AI Token Usage Across 20+ Providers with a macOS Menu Bar App
The article discusses the author's development of a macOS menu bar app called TokenBar to track AI token usage across multiple providers, addressing the problem of scattered dashboards and surprise billing.
Why it matters
TokenBar addresses a common pain point for developers working with multiple AI services, providing a unified view of their token usage and helping them manage their consumption more effectively.
Key Points
- 1TokenBar is a macOS menu bar app that tracks AI token usage across 20+ providers
- 2It polls usage data directly from each provider using CLI tokens, OAuth, API keys, and local file probes
- 3The app is designed to be ambient and privacy-focused, with all data processing happening locally
- 4Using TokenBar helped the author pace their AI usage and avoid unexpected rate limits and billing surprises
Details
The author built TokenBar to address the problem of juggling multiple paid AI services, each with their own rate limits, billing cycles, and usage caps, and having no centralized visibility into their overall consumption. The app sits quietly in the macOS menu bar, showing real-time token usage across providers like Claude, OpenAI, Cursor, GitHub Copilot, and more. It polls the usage data directly from each provider using various authentication methods, without any data leaving the user's machine. The key design decisions were to make the app ambient and privacy-focused, with no account creation or cloud sync required. After using TokenBar, the author found that they started naturally pacing their AI usage and making smarter choices about which tool to use for each task, removing a consistent source of friction from their workflow.
No comments yet
Be the first to comment