ChatGPT's Linux app arrives — minus the feature that matters most

OpenAI shipped a Linux desktop app on Tuesday, finally giving penguin enthusiasts a first-party way to run ChatGPT, ChatGPT Work, and Codex outside a browser tab. The preview supports Ubuntu 24.04 and 26.04 LTS, Debian 13, and Fedora 43 and 44, with .deb and .rpm packages covering both x64 and ARM64 architectures. Voice control, Chrome extensions, and the built-in browser made the cut.

Here’s the catch: native computer use isn’t there. The marquee feature that lets ChatGPT control desktop applications like GIMP or OpenOffice — the one OpenAI has been demoing for months — won’t work on Linux at launch. Appshots, Record & Replay, and voice commands for desktop apps are also missing, an OpenAI spokesperson confirmed.

The timeline tells a story. Mac and Windows users got their desktop apps back in July, and Linux developers have had the Codex CLI and an IDE extension for a while. So what does this app actually add? Mostly a visual interface for juggling multiple projects, files, and code changes. That’s genuinely useful for developers who prefer a GUI over a terminal, but it’s not exactly a headline feature. If you live in the command line, the CLI remains the faster, more direct tool — no GUI wrapper needed.

Missing native computer use on Linux is a notable gap, especially since OpenAI’s agentic push has been the centerpiece of its 2025 roadmap. Linux is the default environment for a huge share of developers and automation workloads, so leaving out the feature that would actually let the agent do things on a user’s behalf feels like shipping the shell without the engine. Whether that’s a technical hurdle, a security consideration, or just a preview-phase limitation, OpenAI hasn’t said. For now, Linux users get a solid chat client and a project dashboard. The agent part? Still waiting.