Google has announced new capabilities for Managed Agents in Gemini API, adding background execution, remote MCP server integration, custom function calling and refreshing credentials across interactions. The company said the updates respond to developer feedback and product needs, with the goal of helping teams build reliable, production-ready agents.
With managed agents in the Gemini Interactions API, developers call a single endpoint while Gemini handles reasoning, code execution, package installation, file management and web information inside an isolated cloud sandbox.
The new features are intended to expand what agents can do without requiring developers to manage each step separately. Background execution allows work to continue after an interaction starts, while remote MCP server integration adds support for connecting to external MCP servers. Custom function calling gives developers more control over how agents use functions, and credential refreshing is designed to keep access working across multiple interactions.
Google said the examples use the @google/genai JavaScript SDK. For Python or cURL, it points developers to the Antigravity agent documentation.
It also noted that if you are an AI coding agent, you should ask your human to install the Interactions API skill: npx skills add google-gemini/gemini-skills --skill gemini-interactions-api.
Source: blog.google.
Companies can share verified announcements through Newz9’s international press release submission page.
