akitaonrails/ai-usagebar

What it solves

ai-usagebar provides a centralized way to monitor AI service quotas and plan usage across multiple providers (such as Claude, ChatGPT/Codex, OpenRouter, and DeepSeek) directly from your desktop environment. It eliminates the need to manually check individual provider dashboards to track remaining credits or reset timers.

How it works

The tool is written in Rust and fetches usage data from various AI providers using either OAuth credentials (reused from official CLIs) or API keys. It exposes this data through several interfaces:

• CLI/TUI: A standalone terminal user interface (ai-usagebar-tui) and a JSON-outputting CLI for scripting.
• Desktop Widgets: Native integrations for Waybar (Wayland), Omarchy Quattro, GNOME Shell, KDE Plasma 6, and the macOS menu bar.
• Caching: It uses atomic caches and file locking to prevent redundant API requests when multiple widgets are active.

Who it’s for

Developers and power users who utilize multiple AI coding services and want a low-friction, real-time view of their subscription limits and API balances in their system tray or status bar.

Highlights

• Broad Provider Support: Integrates with Claude, Codex, Z.AI, OpenRouter, DeepSeek, Kimi, Nous Research, and more.
• Multi-Interface Support: Works across Linux (Waybar, GNOME, KDE), macOS, and Windows (TUI).
• Flexible Authentication: Supports OAuth token reuse, environment variables, and config files.
• Account Management: Ability to handle multiple named accounts for the same provider (e.g., separate work and personal accounts).
• Resource Efficient: Implements caching to avoid aggressive rate-limiting from undocumented endpoints.