Expanding Managed Agents in Gemini API: background tasks, remote MCP and more

2026-07-07 · Source: The Keyword · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, medium

Summary

Google DeepMind announced significant enhancements to Managed Agents in the Gemini API on July 7, 2026, designed to help developers build reliable, production-ready agents. These new capabilities include background execution, allowing asynchronous processing of long-running tasks by returning an interaction ID for status polling. Developers can now integrate remote Model Context Protocol (MCP) servers, enabling direct agent access to private databases and internal APIs from within the secure cloud sandbox. The update also introduces custom function calling, which allows local execution of specific business logic alongside built-in sandbox tools. Furthermore, network credential refresh functionality permits updating access tokens and API keys for existing environments without disrupting the agent's filesystem state or installed packages. These additions transform managed agents into asynchronous workers operating within robust development environments.

Key takeaway

For AI Engineers building production-ready agents with the Gemini API, these updates significantly enhance agent robustness and integration. You can now offload long-running tasks to background execution, ensuring application responsiveness. Integrate your private databases directly via remote MCP servers, expanding agent capabilities securely. Leverage custom function calling for local business logic and refresh network credentials without losing agent state, streamlining your development workflow for complex, reliable AI applications.

Key insights

New Gemini API agent capabilities enable robust, production-ready autonomous agents with enhanced asynchronous, integration, and security features.

Principles

• Agent reliability benefits from async operations.
• Secure sandboxes can extend to private endpoints.
• Modular tool integration supports diverse agent tasks.

In practice

• Use "background: true" for long-running tasks.
• Connect agents to internal APIs via MCP servers.
• Define custom functions for client-side logic.

Topics

• Gemini API
• Managed Agents
• Asynchronous Execution
• Model Context Protocol
• Custom Function Calling
• Credential Refresh

Code references

• googleapis/js-genai

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• How Gemini Managed Agents Works under the Hood — Gemini Managed Agents abstract complex multi-step AI tasks into a single API call, orchestrating an isolated execution environment.
• Introducing computer use in Gemini 3.5 Flash — Gemini 3.5 Flash now natively supports "computer use" for building advanced, cross-platform automation agents with enhanced safety features.
• Gemini Managed Agents: Developer Guide — Gemini Managed Agents provide a serverless platform for deploying AI agents with code execution and secure sandbox capabilities via a single API.
• Subagents have arrived in Gemini CLI — Gemini CLI subagents delegate complex tasks to specialized, isolated agents, preserving primary context and enabling parallel execution.
• Build an AI Research Agent with Google Interactions API & Gemini 3 — Google's Interactions API enables stateful, multi-phase AI agents by managing server-side context and background execution.
• Everything AI from Google IO 2026 in 5 mins! — Google is rapidly expanding its multimodal AI capabilities and developer ecosystem, focusing on cost-efficiency and secure agent deployment.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Keyword.