[BUG] HTTP 400 "Invalid project resource name projects/" — AI Pro subscriber, affects all models, only one account

Google Antigravity
1

I’m a Google AI Pro subscriber experiencing the persistent HTTP 400 error in Antigravity. The issue only occurs on my Pro account — my other Google accounts work perfectly fine with Antigravity, which confirms this is an account-level provisioning issue on Google’s backend.

Environment:

  • OS: Windows
  • Antigravity version: Latest (1.20.3)
  • Account: Google AI Pro subscriber
  • Models tested: All available models (Gemini 3.1 Pro, Gemini Flash, Claude Opus 4.6, etc.) — same error on every model

Error details:
Trajectory ID: 2e38f7e9-1d41-444c-aaa1-4b4b430…
TraceID: 0xdb5d415d19e…
Error: HTTP 400 Bad Request
Status: INVALID_ARGUMENT
Message: Invalid project resource name projects/

What I’ve tried (none worked):

  • Setting GOOGLE_CLOUD_PROJECT environment variable manually with my project ID (api-project-64123584640)
  • Launching Antigravity from terminal with the env variable set
  • Signing out and back in
  • Tested with every available AI model — all fail with the same error
  • The app continues to send requests with an empty project name (projects/) regardless of local configuration

Key observation: Other Google accounts work fine on the same machine and same Antigravity installation. Only my Pro account fails. This points to a server-side provisioning issue where my Pro account was never assigned a valid project ID, or it became unlinked.

Please investigate the account provisioning for my subscription. I have a free 1-year AI Pro subscription through the Google student program and I cannot use the service.

[BUG] HTTP 400 “Invalid project resource name projects/” — Suscriptor AI Pro, afecta a todos los modelos, solo una cuenta

Soy suscriptor de Google AI Pro y tengo el error persistente HTTP 400 en Antigravity. El problema solo ocurre con mi cuenta Pro — mis otras cuentas de Google funcionan perfectamente con Antigravity, lo que confirma que es un problema de provisionamiento a nivel de cuenta en el backend de Google.

Entorno:

  • SO: Windows
  • Versión de Antigravity: Última (1.20.3)
  • Cuenta: Suscriptor Google AI Pro
  • Modelos probados: Todos los disponibles (Gemini 3.1 Pro, Gemini Flash, Claude Opus 4.6, etc.) — mismo error en todos los modelos

Detalles del error:
Trajectory ID: 2e38f7e9-1d41-444c-aaa1-4b4b430…
TraceID: 0xdb5d415d19e…
Error: HTTP 400 Bad Request
Status: INVALID_ARGUMENT
Message: Invalid project resource name projects/

Lo que he intentado (nada funcionó):

  • Configurar la variable de entorno GOOGLE_CLOUD_PROJECT manualmente con mi project ID (api-project-64123584640)
  • Lanzar Antigravity desde la terminal con la variable configurada
  • Cerrar sesión y volver a iniciarla
  • Probado con todos los modelos de IA disponibles — todos fallan con el mismo error
  • La app sigue enviando peticiones con el nombre de proyecto vacío (projects/) sin importar la configuración local

Observación clave: Otras cuentas de Google funcionan bien en la misma máquina y la misma instalación de Antigravity. Solo mi cuenta Pro falla. Esto apunta a un problema de provisionamiento del lado del servidor donde a mi cuenta Pro nunca se le asignó un project ID válido, o se desvinculó.

Por favor investiguen el provisionamiento de mi cuenta. Tengo una suscripción AI Pro gratuita de 1 año por el programa de estudiantes de Google y no puedo usar el servicio

2

Hi @calpee
Welcome to the Google AI Forum!!!

Thank you for bringing this to our attention. We have escalated the issue to our internal teams for a thorough investigation.

To ensure our engineering team can investigate and resolve these issues effectively, we highly recommend filing bug reports directly through the Antigravity in-app feedback tool. You can do this by navigating to the top-right corner of the interface, clicking the Feedback icon, and selecting Report Issue.

3

Thank you for escalating it. As I mentioned, I already filed a bug report through the in-app feedback tool (Report Issue) about a month ago, including all the error details and trace IDs. Unfortunately, I never received any response or follow-up from the engineering team.

I will submit another report through the in-app tool now with updated trace IDs, but I wanted to flag that the previous report went unanswered. Is there a way to track the status of these reports or get a reference/ticket number?