This was resolved for me when I reinstalled Antigravity
This works for macs too. And the message from Nidhin_george_m should be pinned to thee very top!
Update fixed the problem
I have more than 15 Antigravity Ultra and Pro plans, and through various experiments I have inferred what may be causing this error.
First, this error seems to occur on accounts that have received or are currently receiving promotional benefits, such as Ultra discounts, free trials, or free student access.
Second, if there is a discrepancy between the VPN location, the Google login location, and the location where Antigravity is being used, Google seems to detect this and lower the usage allowance. In practice, this error occurred more often on accounts where the VPN, proxy, or desktop address had changed.
Third, three clean accounts that had never received any promotional benefits and had never used a VPN worked normally without any issues.
Fourth, when I tested affected accounts using other countries, there were cases where the error stopped occurring. The error frequency was especially high with U.S. and India IPs.
Conclusion: Based on my inference, this error seems to be triggered more easily when Google considers an account to be abnormal use, such as multiple accounts, VPN usage, promotional abuse, or generally low account reputation, or when the account is being used from countries with heavy traffic. It seems that Google is trying to detect abuse, but in the process it is also causing many errors for legitimate accounts.
2 Likes
Yes, my problem is solved too. I updated to version 1.22.2 and the problem went away, thank you.
1 Like
It’s best to stop paying for this . I’ve canceled it and requested compensation according to the agreed terms.
The same as everyone else. Our servers are experiencing high traffic right now, please try again in a minute.
[URGENT BUG REPORT - Antigravity Remote SSH Server]
Symptom: Since this morning, all Agent threads using Remote SSH have been throwing static 503 errors after completing the first Tool Call (introducing the Tool Output context into the loop causes gRPC interruption). Local execution is normal.
Error: UNAVAILABLE (code 503): No capacity available…
Debug Triage: The error persists even after switching the LLM Model or rolling back the Client-side locally. The Backend Team is requested to review the Payload Packet/gRPC processing engine originating from the antigravity server on the Linux Host; it’s highly likely that this morning’s update is corrupted or overloading the Token structure, causing the LLM Provider Gateway to throw 503 requests.
I’m an AI Ultra subscriber. Honestly, if you’re charging people for this service, shouldn’t you at least send an apology or give us more data? We’re paying customers, after all!
2 Likes
Same issue here, running an Ultra account and all models failing. And this is not a one-off case, but a daily situation. Shame of a service.
1 Like
I have This :
Antigravity Version: 1.22.2
VSCode OSS Version: 1.107.0 (user setup)
Commit: 62335c71d47037adf0a8de54e250bb8ea6016b15
Date: 2026-04-02T19:49:45.095Z (6 days ago)
Electron: 39.2.3
Chromium: 142.0.7444.175
Node.js: 22.21.1
V8: 14.2.231.21-electron.0
OS: Windows_NT x64 10.0.26200
Language Server CL: 893628030
But Still Have Same Issue
yes I’m experiencing the same. using the latest AG v1.22.2 and still facing high traffic issue. so version update did not solve this. it is not a technical issue. it is a global compute resource issue. the autorun in the latest version update helps so it keeps trying until finally completing the task at hand.
2 Likes
I’ve been experiencing the same thing as well. Granted, I am not an Ultra user, but I still have 100% token quota for the models. I was able to actually get some things done earlier this morning; however, I inevitably got the error. Very frustrating 
but ig we’ll manage.
1 Like
Same problem here, constantly getting this annoying error despite I have the ultra plan. This is unacceptable
same erorr i dont know how google has this service
I’ve been dealing with this for two weeks now. I have the $270 Ultra plan, and I have to click “continue” about 30 times before a single task will run. My advice is, don’t waste your money on Antigravity.
1 Like
Implementación de Estrategias de “Backoff” Exponencial
En lugar de reintentos lineales que saturan el motor, se debe implementar un algoritmo de Backoff Exponencial con Jitter. Esto ajusta el tiempo de espera de forma dinámica: si el servidor falla, el equipo espera un tiempo t, luego 2t, luego 4t, añadiendo una pequeña variación aleatoria (jitter) para evitar colisiones de tráfico masivo sincronizado.
2. Reajuste de Carga: Load Balancing & Throttling
Para que el tiempo “gire” a favor del procesamiento del equipo, es necesario optimizar cómo se distribuyen las peticiones:
-
Rate Limiting Local: Controlar la cadencia de salida de las peticiones desde el equipo para no alcanzar los umbrales de seguridad del servidor.
-
Concurrency Control: Limitar el número de hilos de ejecución activos. A veces, procesar menos tareas en paralelo pero con mayor prioridad de CPU reduce la latencia total.
3. Optimización del Stack de Red
Al igual que en la mecánica automotriz, hay que reducir la fricción:
-
Keep-Alive & Connection Pooling: Mantener las conexiones abiertas para evitar el costo de tiempo del “handshake” TCP/TLS en cada petición.
-
Edge Computing: Si la arquitectura lo permite, desplazar parte del procesamiento pesado a nodos más cercanos al usuario para reducir el viaje de los datos.
Hola, comunidad. Me ha pasado en varias oportunidades y he intentado varias cosas, entre ellas la de actualizar y seguía arrojándome error hasta que hice estos pasos y lo solucioné. Espero les ayude:
- Reiniciar el proceso de configuración (Onboarding)
Esta es la solución que ayuda a la mayoría de las personas, ya que obliga a la aplicación a reconstruir su configuración interna.
-
Presiona
Ctrl + Shift + Pen tu teclado. -
Escribe “reset onboarding” y selecciona la opción de Anti-gravity.
-
Sigue los pasos para configurar la herramienta de nuevo e inicia sesión con tu cuenta de Google.
2. Revisar los servidores MCP
Un servidor MCP mal configurado o desactualizado es una de las causas principales de que el “agente” se cierre.
-
Presiona
Ctrl + Shift + Py busca “manage MCP servers”. -
Desactiva todos los servidores MCP y reinicia la aplicación.
-
Actívalos uno por uno. Si el error aparece justo después de activar uno específico, habrás encontrado el problema.
3. Desactivar extensiones conflictivas
Las extensiones antiguas pueden no ser compatibles con las últimas actualizaciones.
-
Usa
Ctrl + Shift + P, escribe “extensions” y selecciona “disable all installed extensions” (desactivar todas). -
Reinicia la aplicación y ve activándolas una a una hasta identificar cuál causa el fallo.
4. Limpiar carpetas de configuración (Solución avanzada)
Si lo anterior falla, el video sugiere borrar los archivos temporales y reinstalar.
-
Cierra la aplicación por completo.
-
Busca en tu computadora la carpeta
%appdata%y elimina las carpetas llamadas “anti-gravity”. -
También revisa en
%localappdata%para borrar rastros similares. -
Desinstala la aplicación desde los ajustes de tu sistema y vuelve a descargarla e instalarla desde el sitio oficial de Google.
1 Like
this issue is a ragebait retrying again and again hits the rate limit and atleast if nothing ultra plan users should not face this issues
Same issue, unable to use. Here is the debug info may help.
Trajectory ID: 837cabf8-e32a-4bfe-8e8a-
Error: HTTP 503 Service Unavailable
Sherlog:
TraceID: 0x6b55fa0f
Headers: {“Alt-Svc”:[“h3=“:443”; ma=2592000,h3-29=“:443”; ma=2592000”],“Content-Length”:[“429”],“Content-Type”:[“text/event-stream”],“Date”:[“Fri, 10 Apr 2026 06:14:08 GMT”],“Server”:[“ESF”],“Server-Timing”:[“gfet4t7; dur=912”],“Vary”:[“Origin”,“X-Origin”,“Referer”],“X-Cloudaicompanion-Trace-Id”:[“6b55fa0f27df2c0d”],“X-Content-Type-Options”:[“nosniff”],“X-Frame-Options”:[“SAMEORIGIN”],“X-Xss-Protection”:[“0”]}
{
“error”: {
“code”: 503,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “claude-opus-4-6-thinking”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
}
],
“message”: “No capacity available for model claude-opus-4-6-thinking on the server”,
“status”: “UNAVAILABLE”
}
}