[bug] Antigravity CLI 1.2.11: HTTP 400 "User location is not support" & HTTP 429 RESOURCE_EXHAUSTED despite 100% quota (US account)

Google Antigravity
,
1

Hi Antigravity Team and Community,

I am encountering a persistent backend error that prevents the Antigravity CLI (v1.2.11) from making any model requests. I am using a Google AI Pro consumer account in the United States. My Google account country is the United States, and my public IP is also located in the United States.

Error Details:

The default backend returns:400

When I test the production Cloud Code backend (https://cloudcode-pa.googleapis.com), the error changes to:RESOURCE_EXHAUSTED (code 429)

Tested Models & Quota:
This happens across all tested models, despite /usage showing 100% quota remaining for Gemini and other models:

屏幕截图 2026-09-25 155840
屏幕截图 2026-09-25 1558401115×628 26.2 KB

  • Gemini 3.8 Flash

  • Gemini 3.1 Pro

  • Claude Opus 4.6

  • GPT-OSS 120B

Troubleshooting & Feedback Submitted:
I have already submitted feedback via the Antigravity CLI. The Error IDs for this issue are:

Error ID: d4b9755f-1892-4c10-8ead-7757f02e7bb5-1

Error ID: e8a56c64-573c-45e0-a29c-f84476c8e08a-9

I also noticed that my situation closely resembles GitHub Issue #1016, where users are experiencing a 400 region error on the default backend and a 429 error on the production backend despite having quota available.

Request:
Could a Google engineer please check the backend trace logs for these Error IDs and verify whether there is an Antigravity / Cloud Code eligibility, entitlement, or account provisioning issue? It appears to be a backend state desynchronization rather than an actual quota exhaustion.

Thank you for your time and assistance.

2

Hi @q_sr,

Thank you for reporting this issue and sharing your environment details, error IDs, and troubleshooting observations. Since you have already tested across multiple models (Gemini 3.8 Flash, Gemini 3.1 Pro, Claude Opus 4.6, GPT-OSS 120B) and confirmed 100% remaining quota via `/usage`, this behavior is caused by backend endpoint routing desynchronization and cached entitlement token mismatches rather than an actual quota exhaustion.

To resolve this issue, please follow these steps:

  • Clear Custom Endpoint Overrides: Check your terminal shell environment variables for any custom endpoint or proxy configurations (such as `ANTIGRAVITY_ENDPOINT` or `CLOUDCODE_ENDPOINT`) and unset them so the CLI routes traffic directly to standard production endpoints (`cloudcode-pa.googleapis.com`).
  • Reset OAuth Credentials & Re-authenticate: Clear your cached authentication tokens by removing the credentials cache in your user configuration directory (e.g. `~/.config/antigravity` or running `agy auth logout`). Then, run `agy auth login` to re-authenticate and fetch a fresh production entitlement token.
  • Verify Account Regional Association: Ensure your Google Account’s registered country is aligned with a supported region by reviewing the Google Terms of Service page (https://policies.google.com/terms) and geographical availability in the Antigravity FAQ (FAQ | Google Antigravity Docs).