CodexBar: An Open-Source Tool to View AI Coding Assistant Usage Stats Without Logging In

Project Overview

As AI coding assistants become increasingly prevalent, OpenAI Codex and Claude Code have become indispensable tools in many developers' daily workflows. However, tracking usage and consumption for these tools typically requires logging into their respective platforms — a tedious process. An open-source project on GitHub called CodexBar is solving this pain point by letting you view AI coding tool usage statistics without logging in.

The project was created by well-known iOS developer steipete (Peter Steinberger), built with Swift, and quickly garnered over 11,700 Stars and 905 Forks after launch — impressive traction by any measure.

What Problem Does It Solve?

Developer Usage Anxiety

For developers who heavily rely on AI coding assistants, understanding their API call volume, token consumption, and cost expenditure is a genuine need. Currently, viewing this data typically requires:

Opening a browser and logging into the OpenAI or Anthropic dashboard
Navigating through multiple menu layers to find the usage statistics page
Manually comparing consumption across different time periods

While this process isn't overly complicated, for users who need to check frequently, having to interrupt their workflow and switch to a browser each time is far from ideal.

CodexBar's Solution

CodexBar takes the form of a macOS menu bar app (as the "Bar" in its name suggests), displaying usage statistics directly in the system menu bar. This means developers can get key usage data with just a glance or a single click, without ever leaving their current development environment.

More importantly, the project emphasizes its "without having to login" feature — viewing statistics without authentication. This is likely achieved by reading local API key configuration files or cached data, eliminating the need for additional authentication flows.

Technical Highlights

Native Swift Development

The project is written in Swift, making it a native macOS application. Compared to Electron or web-based solutions, a native Swift app offers several advantages:

Lower resource consumption: As a menu bar resident app, being lightweight is crucial
Better system integration: Full access to macOS native APIs and UI components
Smoother user experience: Fast response times with visual styling consistent with the system

Multi-Platform Support

CodexBar supports usage statistics for both OpenAI Codex and Claude Code — the two major AI coding assistants. This multi-platform aggregation design is highly practical. Many developers use multiple AI tools simultaneously, and being able to view usage for all tools in a unified interface significantly boosts efficiency.

Community Response

Over 11,000 Stars is an impressive achievement for a utility project, reflecting several trends:

The user base for AI coding tools is already massive, and ecosystem demands around these tools are exploding
Developer need for usage management is real and urgent, especially under pay-per-use pricing models
There's still a huge market for "small but beautiful" developer tool projects — solving one specific pain point can earn widespread recognition

The 905 Forks also indicate that many developers want to customize or contribute code on top of this foundation, showing community validation of the project's extensibility.

Who Is It For?

macOS developers who use OpenAI Codex or Claude Code daily
Team leads who need to monitor AI tool usage costs
Technical professionals looking to optimize their AI tool usage efficiency
Developers interested in the AI coding assistant ecosystem

Conclusion

CodexBar is a textbook example of turning an itch into a product within the AI coding tool ecosystem. It doesn't attempt complex features — it focuses solely on solving one specific problem: making it easier for developers to understand their AI tool usage. In an era where AI coding assistants are becoming standard equipment, the value of such auxiliary tools will only continue to grow. If you're a macOS user who frequently uses AI coding assistants, CodexBar is worth a try.