Skip to content
4ALL API Docs
Main Site News Console

and more" or "etc

· Google AI Translated
GoogleGemini

July 7, 2026

We are adding support for new capabilities, such as background execution for asynchronous interactions, easy connection to remote MCP servers, custom functions, and credential refreshing.

Philipp Schmid

Developer Relations Engineer, Google DeepMind

Mariano Cocirio

Product Manager, Google DeepMind

Managed agents feature bundle launch

Listen to article

[[duration]] minutes

This article was generated by Google AI. Generative AI is still experimental.

Today, we are announcing new capabilities for Managed Agents in the Gemini API, including background execution, remote MCP server integration, custom function calling, and credential refreshing across interactions. These updates directly address developer feedback and product requirements, helping you build reliable, production-ready agents.

When using Managed Agents in the Gemini Interactions API, you only need to call a single endpoint, and Gemini handles reasoning, code execution, package installation, file management, and network information retrieval in an isolated cloud sandbox.

If you are an AI coding agent, please ask your users to install the Interactions API skill: npx skills add google-gemini/gemini-skills --skill gemini-interactions-api.

Below is an example using the @google/genai JavaScript SDK. For Python or cURL, please check the Antigravity agent documentation.

Building Autonomous Agents with Scaling Capabilities

Long-Running Background Execution

For long-running tasks, keeping an HTTP connection open is unreliable. Passing background: true runs the interaction asynchronously on the server side. The API immediately returns an ID that the client application can use to poll status, stream progress, or reconnect later when the agent completes remotely. For more details, see the Background Execution Guide.

import { GoogleGenAI } from "@google/genai";


const client = new GoogleGenAI({});


// 1. 在后台启动一项长时间分析
const interaction = await client.interactions.create({
  agent: "antigravity-preview-05-2026",
  input: "Clone https://github.com/googleapis/js-genai, find all TODO comments in the source code, and categorize them by module and priority in a markdown report.",
  environment: "remote",
  background: true,
});


console.log(`后台任务已启动。交互 ID: ${interaction.id}`);


// 2. 异步轮询，不阻塞打开的 HTTP socket
let result = interaction;
while (result.status === "in_progress") {
  await new Promise((resolve) => setTimeout(resolve, 5000));
  result = await client.interactions.get(interaction.id);
}


if (result.status === "completed") {
  console.log("任务完成：\n", result.output_text);
} else {
  console.error(`任务结束，状态为：${result.status}`);
}

Remote MCP Server Integration

You can now connect Managed Agents directly to remote Model Context Protocol (MCP) servers, instead of writing custom agent middleware to access private databases or internal APIs.

You can mix remote tools with built-in sandbox capabilities. Pass the mcp_server tool during the interaction, and use it alongside Google Search or code execution, allowing the agent to communicate with your endpoints within a secure sandbox. As you expand external tools and APIs for your agents, please follow the best practices.

import { GoogleGenAI } from "@google/genai";


const client = new GoogleGenAI({});


const interaction = await client.interactions.create({
  agent: "antigravity-preview-05-2026",
  input: "Check our internal observability server for recent latency spikes in the auth service and correlate them with git commits.",
  environment: "remote",
  tools: [
    { type: "google_search" },
    { type: "code_execution" },
    {
      type: "mcp_server",
      name: "internal_telemetry",
      url: "https://mcp.internal.example.com/mcp",
    },
  ],
});


console.log(interaction.output_text);

Custom Function Calling Alongside Sandbox Tools

You can add custom tools for local execution alongside built-in sandbox tools. The API uses step matching. Built-in tools run automatically on the server side, while custom functions switch the interaction to requires_action, allowing your client to execute local business logic.

import { GoogleGenAI } from "@google/genai";


const client = new GoogleGenAI({});


// 1. 定义一个自定义领域函数
const getWeatherTool = {
  type: "function",
  name: "get_weather",
  description: "Gets the current weather for a given location.",
  parameters: {
    type: "object",
    properties: {
      location: {
        type: "string",
        description: "The city and country, e.g. San Francisco, USA",
      },
    },
    required: ["location"],
  },
};


// 2. 同时使用内置代码 execution 和自定义函数调用智能体
const interaction = await client.interactions.create({
  agent: "antigravity-preview-05-2026",
  input: "Check the weather in Tokyo, write a Python script to convert the temperature to Fahrenheit, and save the result to weather.txt.",
  environment: "remote",
  tools: [
    { type: "code_execution" },
    getWeatherTool,
  ],
});


// 3. 清晰地处理自定义函数执行
if (interaction.status === "requires_action") {
  // 文件系统和沙箱工具会自动执行，并生成匹配的 function_result 步骤。
  // 我们筛选出需要客户端执行的待处理领域调用。
  const executedCalls = new Set(
    interaction.steps
      .filter((s) => s.type === "function_result")
      .map((s) => s.call_id)
  );


  const pendingCalls = interaction.steps.filter(
    (s) => s.type === "function_call" && !executedCalls.has(s.id)
  );


  for (const call of pendingCalls) {
    console.log(`正在执行客户端工具：${call.name}（ID：${call.id}）`);
    // 执行你的本地 API/数据库查询，并在第 2 轮把 function_result 返回
  }
}

Network Credential Refreshing

Access tokens and short-lived API keys expire. You can refresh credentials or rotate keys by passing an existing environment_id along with the new network configuration.