Google Adds Background Execution and Remote MCP Support to Gemini API Managed Agents

Google DeepMind today announced several major updates to Managed Agents within the Gemini Interactions API, including background execution, remote MCP server integration, custom function calling, and network credential refresh. The additions mark a significant evolution from conversational interactions toward production-ready asynchronous agent workflows.

Managed Agents are part of the Gemini Interactions API, where a single endpoint call lets Gemini handle reasoning, code execution, package installation, file management, and web information retrieval inside an isolated cloud sandbox. The new updates directly respond to developer feedback about real-world production requirements.

Background execution is the most critical new feature. Developers can now pass background: true to run interactions asynchronously on the server. The API immediately returns an ID, which client applications can use to poll for status, stream progress, or reconnect later — even if the originating device goes offline. This addresses the fragility of holding HTTP connections open for long-running tasks.

Remote MCP server integration eliminates the need for custom proxy middleware. Developers can connect managed agents directly to remote Model Context Protocol (MCP) servers by passing an mcp_server tool at interaction time, allowing agents to communicate with private databases and internal APIs from their secure sandbox.

Custom function calling enables developers to add custom tools alongside built-in sandbox tools for local execution. Meanwhile, network credential refresh solves the problem of short-lived access tokens and API keys — developers pass their existing environment_id with a new network configuration to refresh credentials or rotate keys on the next interaction.

These updates turn managed agents into asynchronous workers that operate inside real development environments without blocking the application. For developers building long-running data processing, scheduled report generation, or cross-system orchestration tasks, background execution and remote MCP connectivity are two key enablers.

Notably, Google also published an Interactions API skill for AI coding agents to install directly via npx skills add, suggesting Google is pushing toward inter-agent operability.

As the AI agent platform race heats up — Microsoft with Copilot and Foundry Agent Service, Anthropic with Claude Cowork — Google is carving a differentiated path with Gemini Managed Agents, emphasizing sandbox security, open MCP protocols, and asynchronous execution. The key questions ahead are pricing and quota strategy, and whether enterprise customers will move from prototyping to production deployment at scale.