Discover and explore top open-source AI tools and projects—updated daily.
Fei-AwayEnhance AI desktop applications
Top 4.5% on SourcePulse
Summary
Fei-Away/Codex-Dream-Skin provides a local injection tool for the Codex desktop client, enabling users to apply custom themes and backgrounds without altering the official application. It targets Codex users seeking a personalized coding environment and offers a visually enhanced experience.
How It Works
The project employs local CDP (Chrome DevTools Protocol) injection to dynamically apply external themes. It overlays a full-window 16:9 wallpaper and ensures UI elements remain interactive and native. Theme configurations are independent of API settings, and an optional AI model relay is provided for services like Codex, Claude Code, and Grok.
Quick Start & Requirements
Installation scripts are provided for macOS and Windows.
macos/ and run ./scripts/install-dream-skin-macos.sh. Themes can be switched using ~/.codex/codex-dream-skin-studio/scripts/switch-theme-macos.sh..\windows\scripts\install-dream-skin.ps1 followed by .\windows\scripts\start-dream-skin.ps1 via PowerShell.
Requires the Codex desktop client. Detailed platform guides are available in macos/README.md and windows/SKILL.md.Highlighted Details
Maintenance & Community
Contributions are managed via GitHub Issues and Pull Requests, with specific templates provided. Users are encouraged to run verification/restore scripts before submitting changes. Further details are in docs/PROJECT.md.
Licensing & Compatibility
The project is licensed under the MIT License (macos/LICENSE). It is explicitly stated as a non-official OpenAI product. Commercial use or public redistribution of provided assets requires independent verification of portrait, material, and trademark rights.
Limitations & Caveats
This is a third-party tool and not officially endorsed by OpenAI. It does not modify the official Codex binary or signature. Users must independently clear rights for any custom assets used for redistribution.
2 weeks ago
Inactive