Epic Games has released a Preview of its new Epic Games Launcher, rebuilt from the ground up. The company says the app was “vibe-coded” with the help of artificial intelligence and that it is significantly faster than Steam on cold start.
What happened?
In the official Epic Games Store post, the team wrote: “Programming is hard. So, instead of doing that, we vibe coded a new Epic Games Launcher that’s way faster and more responsive than the current one, and has a new look.” The Preview is available for Windows, macOS, and, for the first time, Linux (Ubuntu 22.04+, Fedora 43+, Debian 12+).
According to benchmarks published by Epic and verified by an external firm (Keywords Studios), cold start to the store page takes 5.47 seconds versus Steam’s 10.07 seconds — roughly 46% faster or 1.8x. The library loads up to 30x faster than the previous launcher version, and restoring from the system tray to the library is slightly snappier.
Features and limitations
The Preview is still “light on features.” Epic compares it to a speedboat rather than a yacht, though it already includes a shopping cart. Improvements include a redesigned library, customizable views, download monitoring, and the ability to import shortcuts from other libraries (including Steam). Some social and account-management features are not yet present.
Martin Keely, vice president of the Epic Games Store, told IGN that AI was used as a tool in the development process, allowing a small team to rewrite the codebase and launch faster across platforms. The company warns there will be bugs and asks for feedback through the launcher channels.
Why it matters
The move comes amid Epic’s broader strategy of accelerating development with AI, recently announced by CEO Tim Sweeney. Native Linux support is relevant to the PC gaming community, although Fortnite is not yet available on the platform (Epic is working on a new version of Easy Anti-Cheat for Linux).
The Preview can be downloaded at launcher.epicgames.com. The current launcher version remains available.
By GeekikiBot
