Google Deepmind adds background execution and MCP support to Gemini API managed agents

2026-07-08 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

Google Deepmind has introduced four new features for Managed Agents within the Gemini API, enhancing developer capabilities. As of July 8, 2026, developers can now utilize Background Execution to run agents asynchronously without requiring an open HTTP connection. The update also enables direct connection of remote MCP (Model Context Protocol) servers to internal databases or APIs. Furthermore, developers can integrate custom functions alongside the existing built-in sandbox tools. A final addition allows for credential refreshing, such as tokens, between agent interactions without losing the sandbox's current state. These features are accessible through the Gemini Interactions API, with supporting code examples provided for JavaScript, Python, and cURL in the documentation.

Key takeaway

For AI Engineers developing with the Gemini API, these updates significantly expand agent capabilities and operational flexibility. You can now design more robust applications by utilizing background execution for complex tasks and integrating custom functions directly into your agent workflows. This allows for seamless connection to internal data sources via MCP servers and ensures secure, uninterrupted agent interactions through automatic credential refreshing, streamlining development and deployment of sophisticated AI agents.

Key insights

The Gemini API's Managed Agents now support asynchronous background execution, custom functions, and enhanced data integration.

Principles

• Asynchronous execution improves agent efficiency.
• Custom functions extend agent capabilities.
• Direct data access enhances agent utility.

In practice

• Run agents in background for long tasks.
• Connect agents directly to internal APIs.
• Use custom code with sandbox tools.

Topics

• Gemini API
• Managed Agents
• Background Execution
• Model Context Protocol
• Custom Functions
• AI Agents

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, Software Engineer, Machine Learning 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.
• 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.
• Combine Built-in Tools and Function Calling in the Gemini Interactions API — Gemini API now combines built-in and custom tools, automating orchestration and context circulation for complex agent workflows.
• 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.
• Unleash Your Development Superpowers: Refining the Core Coding Experience — Gemini Code Assist enhances developer productivity through AI-powered automation, collaboration, and customization features.

Open in AIssential →

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