I’m running into an issue with my deployment and could use some insight. I’m currently building a web app integrated with the Google AI Studio (Gemini) API. Everything works flawlessly in my preview—the API key is read correctly, and the model functions exactly as expected.
However, the moment I publish the app to Google Cloud Run, the API requests fail completely.
Are there any specific Cloud Run configurations (like handling environment variables in secrets, egress network settings, or API key restrictions) that I might have missed?
Thanks in advance for any pointers!