
Deploy personal AI agents with Cloud Run instances
Google introduced Cloud Run instances, providing dedicated, singleton compute runtimes designed for long-lived, stateful workloads such as personal AI agents.
Why it matters
Developers can now run continuous AI agents without the high cost of dedicated virtual machines or the scaling limitations of stateless services.
The details
Each instance receives a persistent HTTPS URL that remains unchanged across restarts. The service is currently in preview and utilizes shared vCPU with burst budgets for predictable pricing. Users have the ability to stop and resume instances as needed.
What's next
Google will soon launch SSH access for both Cloud Run instances and Cloud Run services.
Get the weekly recap
The stories like this one, picked and explained — once a week, straight to your inbox.