ANTIGRAVITY LABJP
Articles/Antigravity Basics
Antigravity Basics/2026-04-26Beginner

Configuring Antigravity's Display Language — Separating UI, AI Responses, and Code Comments

How to switch Antigravity's display language and tune the agent's response language separately. Walks through the three independent layers of language settings, with the gotchas you only discover after running into them.

Antigravity356Localizationi18n3Language SettingsSetup5UI2

import { Callout } from '@/components/ui/callout';

When you first open Antigravity, the menus and error messages all appear in English. If you want to switch the UI to Japanese, German, or any other language, the path through the settings panel is not as obvious as it could be. This guide walks through it with the gotchas I have actually run into.

The key insight is that Antigravity's language settings are split into three independent layers: the UI itself, the AI agent's response language, and code comment language. Without knowing this, you can easily end up in a half-translated state where the menus are localized but the AI keeps replying in English. We will set them in order.

Layer 1: UI Display Language

This is the menu bar, notifications, and settings panel.

  1. Launch Antigravity, open Settings via the gear icon in the status bar (or Ctrl+, / Cmd+,)
  2. Search for display language in the settings search box
  3. Set Display Language to your target locale (for example, 日本語 (ja), Deutsch (de), Français (fr))
  4. When prompted to restart, click Restart

After restart, the menu bar should be localized. If it is not, fully quit Antigravity (including background processes) and relaunch — that resolves cases where the language pack downloaded correctly but did not get picked up.

ℹ️
You can also run `Configure Display Language` from the command palette (`Ctrl+Shift+P` / `Cmd+Shift+P`). If the language pack is not yet installed, Antigravity downloads it automatically.

Layer 2: AI Agent Response Language

A localized UI does not change how the agent talks. You will still get English answers unless you tell it otherwise. There are two ways to handle this.

Option A: Set a global default

Settings → AI Assistant → Default Response Language. Pick the target language and every new chat starts there.

Option B: Set the policy in agents.md per project

Drop an agents.md (or .cursorrules, or AGENTS.md) at the project root and declare your language policy:

# Agent Rules
 
## Language Policy
- Respond to the user in Japanese (politeness register)
- Keep code comments in English unless explicitly told otherwise
- Error explanations: Japanese
- Documentation: match the file's existing language

I prefer Option B for one practical reason: if you also work on English-only projects, the global setting forces you to flip back and forth. A per-project agents.md makes the policy travel with the codebase.

Layer 3: Code Comment Language

This one is sneaky. The agent might respond in your localized language but generate code comments in English (or vice versa), which clashes with team conventions.

Be explicit in agents.md:

## Code Comments
- Domestic-only project: comments in the local language
- International team: comments in English
- Code intended for OSS release: always English

For the Dolice Labs sites I run, the policy is "UI in Japanese and English, code comments in English, agent explanations in Japanese." That keeps the codebase friendly for global readers while letting me operate the project in my own language.

Common Snags

A few patterns I have seen repeatedly.

"Failed to download language pack"

Almost always TLS interception inside a corporate network. Set the CA bundle and restart:

export NODE_EXTRA_CA_CERTS=/path/to/your/corporate-ca-bundle.pem

Your IT team can supply the path.

UI is localized, but the command palette stays in English

A leftover cache. Delete %APPDATA%/Antigravity/Cache on Windows or ~/Library/Caches/com.google.Antigravity on macOS, then relaunch.

Agent occasionally drifts back to English

Usually the policy in agents.md has been pushed out of context as the session grew. State the language preference once at the start of any long session, or remind the agent partway through. A more robust fix is to keep agents.md short and high-priority enough that it survives compaction.

Some keyboard shortcuts changed

Localized builds occasionally remap shortcuts to better fit local keyboard layouts. If a familiar shortcut stops working, override it explicitly in Settings → Keyboard Shortcuts.

A "Sweet Spot" Configuration

Here is the configuration I actually run, designed for a solo developer who ships globally:

UI Language:           Local (e.g. Japanese)
AI Response Language:  Local, declared in agents.md
Code Comments:         English (assume OSS release)
Documentation:         Both local and English (bilingual articles)
Commit Messages:       English
PR / Issue Titles:     English

This setup minimizes day-to-day friction (you work in your native language) while keeping the artifacts portable enough for an international audience.

Next Step

Start with Layer 1 — the UI switch is a five-minute change with a big quality-of-life return. Once that feels right, write a short agents.md that captures your language policy. After a few projects, you will start to develop a personal style around how detailed those policies should be.

Share

Thank You for Reading

Antigravity Lab is ad-free, supported entirely by members like you. We publish practical guides daily with implementation code, benchmarks, and production-ready patterns. If you've found it useful, we'd love to have you on board.

  • Copy-paste ready implementation code
  • New advanced guides published daily
  • $5/mo or $15 for lifetime access
View Membership →

If you found this article helpful, a small tip ($1.50) would mean a lot to us. Your support helps keep this site ad-free and covers server and hosting costs.

Related Articles

Antigravity2026-08-16
Running Antigravity CLI 1.1.13 on a Machine You Cannot Sign In From
I put agy on a box with no browser and the sign-in screen stopped me cold. Here is how direct GEMINI_API_KEY auth from CLI 1.1.13 works, how to confirm it took effect, and why you should not hand it the key your app already uses.
Antigravity2026-08-10
Guarding the Turn Boundary in Non-Interactive Runs: A Command Table and a Fail-Closed Gate
A scheduled job meant only to record usage was calling the model on every run. Here is how I split commands into ones that start a turn and ones that do not, built a gate that refuses anything unclassified, and measured it against real scripts.
Antigravity2026-08-02
I Copied the Same agent.md Into Another Repo and It Quietly Did a Different Job
CLI 1.1.6 lets you carry agent definitions around as files. I dropped one definition into eight repos, built a preflight that resolves its declared capabilities before the agent runs, and measured it against a naive checker.
📚RECOMMENDED BOOKS
Build a Large Language Model (From Scratch)
Sebastian Raschka
LLM Dev
Prompt Engineering for LLMs
Berryman & Ziegler
Prompting
AI Engineering
Chip Huyen
AI Eng
* Contains affiliate links
See all →