Google used I/O 2026 to add new Gemini-based developer tools across AI Studio and the Gemini API. The updates include native Android app generation and managed agents that run in isolated Linux environments.
Google AI Studio now includes tools for app creation, previewing, and testing. The Managed Agents feature can run agents that use tools, execute code, manage files, and browse the web. Google also used I/O to report wider AI usage across its products. CEO Sundar Pichai said Google now processes 3.2 quadrillion AI tokens per month, up from 480 trillion a year earlier.
Android app building in AI Studio
Google AI Studio now supports native Android app building from the build tab. Developers can select “Build an Android app” and use prompts to generate an app. Google said the workflow does not require local software installation or SDK management.
The Android app-building feature generates Kotlin code using Jetpack Compose patterns. Google said the initial Android release is focused on personal utility apps, simple social apps, and AI-powered apps with Gemini API integrations. It also supports hardware-enabled experiences using device features such as camera and GPS.
The workflow includes an embedded Android Emulator, device installation through Android Debug Bridge, and direct publishing to an internal testing track in Google Play Console. Google said AI Studio can connect to a Google Play Developer account for internal testing.
Apps created with AI Studio still need to meet Google Play’s quality and review standards. Google told The Verge that it is not changing its app review process for AI Studio-generated apps.
Google said AI Studio is also being connected with other Google services. Apps built inside AI Studio can now access Google Workspace data, including Sheets, Drive, and documents.
Google announces Antigravity 2.0
AI Studio projects can also be exported to Google Antigravity for local development. Google said the export includes conversation history, project files, and secrets.
Google also announced Antigravity 2.0 at I/O. The company described it as an agent-first development platform. The update includes an updated desktop app, a command-line interface, and an SDK for custom workflows.
Google also said Google Cloud customers will be able to connect to Antigravity to build projects.
Google said AI Studio projects can be exported to Antigravity. Managed Agents is powered by the Antigravity agent. According to Google, Antigravity is intended to manage and deploy agents across AI Studio, Android, Firebase, and Google Cloud.
Managed Agents come to the Gemini API
Google also announced Managed Agents in the Gemini API. The feature allows developers to create an agent with a single API call.
The agent can reason, use tools, and execute code. It runs inside an isolated, temporary Linux environment. The feature is powered by Google’s Antigravity agent, which is built on Gemini 3.5 Flash.
The Managed Agents feature is available in Public Preview through the Interactions API and in Google AI Studio. Google said the feature handles parts of the infrastructure work for production-grade agents, including sandbox setup and environment management.
Each call to the Antigravity agent provisions a remote Linux environment. Within that environment, the agent can plan tasks, call tools, and execute code. It can also manage files and browse the web to retrieve current information. Each interaction can create or receive an environment. Developers can use follow-up calls to resume a session while keeping files and state intact.
Google said developers can define custom agents by extending the Antigravity agent with their own instructions and skills. Developers can write these in markdown files such as AGENTS.md and SKILL.md, then register them as a managed agent.
Google’s Gemini API documentation says developers can customise the Antigravity agent with system instructions and tool settings. They can also mount files such as AGENTS.md and SKILL.md, or save the configuration as a managed agent invoked by ID.
Google cited Deep Research, introduced in December, as its first managed agent for research tasks. Google said it is now opening its agent harness and infrastructure so developers can build custom managed agents.
(Photo by Daniel Romero)
See also: Google says AI helped build zero-day exploit targeting 2FA bypass
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.
Developer Tech News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.