Apply a Chinese localization to the Antigravity code editor so its menus, shortcuts, and AI panels display in simplified Chinese.
Use the conservative core-only patch to translate main menus and shortcuts without touching AI panels for a more stable result.
Restore the original English interface instantly using the provided restore batch file if the patch causes any issues.
This repository provides a Chinese translation patch for the Antigravity 2.0 desktop code editor. Antigravity is a coding environment with AI features built in, this project exists because the editor ships only in English, and many Chinese-speaking users want its menus, settings, and AI panels in their native language. The patch comes in two modes. The first is a basic core translation that covers the main editor menus, keyboard shortcut descriptions, and core extension messages. This mode is conservative and does not touch the AI interface panels, giving a stable result. The second mode goes further and also translates the AI settings pages, navigation panels, and other UI surfaces that are rendered through a custom protocol inside the app. Translating those panels required intercepting the app's internal URL scheme rather than directly replacing files, which is how the project avoids the freezing-on-startup problem that older patching approaches caused. The translation dictionary contains around 300 manually reviewed phrase mappings. The project deliberately skips dynamically generated text such as project names or live AI output, since replacing those unreliably could corrupt the app's displayed text. For everyday users, applying the patch means downloading the ZIP, closing Antigravity, and running a batch file. One batch file applies the full AI-panel translation, another applies only the conservative core translation, and a third restores the original English. For users comfortable with the command line, PowerShell scripts offer the same options with more control. The project uses a version fingerprint system to detect when Antigravity updates its internal bundle. If the detected fingerprint does not match a known-compatible one, the AI-panel translation step falls back gracefully rather than breaking the editor. The code is written in Python and PowerShell, and the project is released under the MIT license.
← kakarotto-baroko on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.