Google Announces New Capabilities for Gemini API Managed Agents
Google has announced the expansion of its managed agent system on the Gemini API with background processing, remote server integration, and customized functions.
In response to developer feedback, Google has added significant capabilities to managed agents within the Gemini API, such as long-running background execution, remote Model Context Protocol server connection, and network credential refresh.
Background Execution Support
Developers can now use the background true parameter in their requests to allow interactions to run asynchronously on the server side.
API returns a unique ID for clients to query the status or track progress.
Remote Server Integrations
Managed agents can connect directly to remote Model Context Protocol servers without needing to write custom middleware code.
Custom Function Calling
While built-in tools run automatically on the server, custom functions alter the interaction state for local execution on the client side.
Credential Refresh
Access tokens and short-lived API keys can be updated or rotated with a new network configuration in subsequent interactions.