Following the preview at I/O 2025, Google is releasing an “Agent Mode” for Gemini in Android Studio.
Agent Mode lets developers accomplish “complex development tasks” like generating unit tests, complex refactors, and “build my project and fix any errors.” Google frames this as delegating “routine, time-consuming work to the agent,” so that developers can focus on “more creative, high-value work.” Other examples include:
- Extract any hardcoded strings used across my project and migrate to strings.xml
- Add support for dark mode to my application
- Given an attached screenshot, implement a new screen in my application using Material 3
You can “describe a complex goal in natural language” and Gemini will formulate a plan that “can span multiple files in your project” and use a “range of IDE tools for reading and modifying code, building the project, searching the codebase and more.”
Developers can “review, refine and guide the agent’s output at every step,” with an “Accept” or “Reject Change” screen. Google will also offer an “Auto-Approve” option to “iterate on ideas as rapidly as possible.”
Agent Mode is available in the latest Android Studio Narwhal Feature Drop Canary release for all users, with business tier subscribers getting access “in the coming days. After launching Gemini from the sidebar, there will be a new Agent tab alongside the existing Chat.
Meanwhile, Agent Mode can take advantage of Gemini 2.5 Pro’s 1 million token context window if you enter your Gemini API key: File (Android Studio on macOS) > Settings > Tools > Gemini. This allows for more instructions, code, attachments, and higher-quality responses.
Note that business tier subscribers already get access to Gemini 2.5 Pro and the expanded context window automatically with their Gemini Code Assist license, so these developers will not see an API key option.
Agent Mode can also interact with external tools, like GitHub to “create pull requests directly from Android Studio,” via the Model Context Protocol (MCP).
For this initial release, we support interacting with external tools via the stdio transport as defined in the MCP specification. We plan to support the full suite of MCP features in upcoming Android Studio releases, including the Streamable HTTP transport, external context resources, and prompt templates.
Google teases more agentic capabilities for Android Studio in the future.
FTC: We use income earning auto affiliate links. More.
Comments