CodexBar is a free and open-source macOS menu bar application developed by Peter Steinberger. It is designed to track and display usage information for a wide array of AI coding providers. The application aims to provide developers with real-time insights into their AI service consumption, including usage windows, credit balances, and reset countdowns, directly from their macOS menu bar, thereby helping them manage their resources more effectively. Supporting over 40 different providers, CodexBar offers a centralized monitoring solution.

This utility's primary functionality is to monitor and present up-to-date data regarding AI coding provider usage. It meticulously tracks session, weekly, and monthly usage windows, offering precise countdowns to upcoming resets. Furthermore, it displays credit balances and estimated monthly spend, enabling users to make informed decisions about their coding tasks. The application provides flexibility, allowing users to view each provider as a distinct status item or consolidate multiple providers into a single, integrated view.

Some of the key features are:

• Provider Tracking: Supports over 40 popular AI coding providers, each displaying individual usage windows, reset countdowns, charts, and dedicated provider menus.
• Resets and Planning: Provides per-provider session, weekly, and monthly usage windows with countdowns to the next reset, assisting users in planning tasks around service limitations.
• Credit and Cost Monitoring: Displays credit balances and monthly spend where such data is exposed by the provider, complemented by configurable local cost scans for specific services like Codex and Claude.
• Status and Incidents: Includes provider status polling, which surfaces incident badges directly in the menu and an indicator overlay on the menu bar icon, alerting users to service issues.
• CLI and Widgets: Offers WidgetKit widgets for supported providers and a bundled codexbar command-line interface (CLI) for scripting and continuous integration (CI) purposes, with builds available for both macOS and Linux.
• Installation and Updates: Available as a universal application via GitHub Releases with built-in Sparkle auto-updates, or installable through Homebrew (brew install --cask steipete/tap/codexbar) for Apple Silicon, with updates managed via brew upgrade.
• Privacy: Reuses existing provider authentication methods, such as OAuth, device flow, API keys, browser cookies, and local files, ensuring that no passwords are stored directly by the application.
• macOS Permissions: Requires specific macOS permissions, including Full Disk Access for accessing Safari cookies, Keychain access for cookie decryption and OAuth flows, and Files & Folders prompts when provider helpers need to read project directories.

To operate CodexBar, users can install it on macOS (version 14 or newer) either by downloading the universal application package from GitHub Releases or by utilizing the Homebrew package manager. Once installed, the application resides in the macOS menu bar. Users then configure the specific AI coding providers they wish to monitor. CodexBar subsequently fetches and displays real-time usage data for these services. The flexible interface allows for individual status items per provider or a consolidated view for multiple providers. Additionally, the integrated CLI offers programmatic access to usage data, expanding its utility for developers.

Some common use cases include:

• Cost Management: Users can closely monitor their expenditure on various AI coding services, helping them adhere to budgets and avoid unexpected charges or overages.
• Usage Planning: Developers can optimize their workflow by understanding when API limits reset, enabling them to strategically schedule computationally intensive tasks or manage daily coding sessions more efficiently.
• Multi-Provider Overview: The application provides a consolidated view of usage statistics across numerous AI coding platforms, simplifying management for developers who leverage multiple services concurrently.
• Automation and Scripting: The bundled codexbar CLI allows for the integration of AI coding usage data into custom scripts or CI/CD pipelines, facilitating automated reporting or triggering actions based on predefined usage limits.