How I Forked OpenUsage and Built a Linux-First Community Edition

Over the last few months, AI coding tools have become a major part of my daily workflow.

Like many developers, I found myself using multiple services at the same time:

• Claude Code
• Codex
• Cursor
• Copilot
• Gemini

The problem wasn't the tools themselves.

The problem was keeping track of usage limits.

I constantly found myself opening different dashboards just to answer one simple question:

How much usage do I actually have left?

Looking for a Solution

I started looking for a lightweight desktop application that could live in the system tray and show usage information for all my AI coding subscriptions in one place.

I eventually found OpenUsage.

The project was open source, well designed, and already supported many providers.

There was only one problem.

I use Linux.

At the time, Linux support was either missing or incomplete.

Why I Created OpenUsage Community

While exploring the project, I learned that the original OpenUsage project was moving toward a Swift and macOS-first architecture.

That made me realize there was room for a community-maintained continuation focused on Linux and cross-platform support.

So I created:

OpenUsage Community

GitHub:

https://github.com/openusage-community/openusage

Goals

The project has a few simple goals:

• First-class Linux support
• Keep the Tauri-based cross-platform architecture
• Support multiple AI coding providers
• Remain open source and community-driven
• Make installation easy with AppImage, .deb and .rpm packages

What Works Today

Current features include:

• Linux system tray integration
• AppImage builds
• Debian / Ubuntu packages
• RPM packages for Fedora and related distributions
• Multiple AI provider integrations
• Local-first architecture
• Automatic updates

The application allows me to monitor usage across providers without opening multiple websites and dashboards throughout the day.

What I Learned

One thing I learned very quickly is that Linux users care deeply about quality.

It isn't enough for an application to compile.

It needs to:

• install cleanly
• integrate with the desktop
• work on GNOME
• work on KDE
• behave correctly on Wayland

This pushed me to start building a more serious testing and CI/CD strategy around the project.

What's Next

The next areas I'm focusing on are:

• stronger Linux desktop integration
• automated GUI testing
• screenshot regression testing
• improved packaging
• community contributions

I'm also interested in hearing how other developers keep track of AI coding subscription usage.

Do you monitor usage limits manually?

Do you use dashboards?

Or do you simply wait until the limit is reached?

Feedback Welcome

If you're a Linux user, I'd love to hear your thoughts.

If you want to try the project or contribute:

https://github.com/openusage-community/openusage

Feedback, bug reports, and pull requests are always welcome.