product updateGoogle

Google opens Android Studio to more than 40 third-party coding agents with Bring Your Own Agent

TL;DR

Google is opening Android Studio to more than 40 external coding agents, including Anthropic's Claude Agent and OpenAI's Codex, through a new Bring Your Own Agent feature in Android Studio Rabbit 2. At the Android Dev Summit, it also announced an alpha Agentic API powered by Gemini Nano 4, a beta AppFunctions API, and the $1,099 XREAL AURA smart glasses.

4 min read
0

Google is opening Android Studio to more than 40 third-party coding agents. Android Studio Rabbit 2 introduces a "Bring Your Own Agent" (BYOA) feature built on the open Agent Client Protocol (ACP), announced at the Android Dev Summit on October 28, 2026.

Agents inside the IDE

Developers can install and run Anthropic's Claude Agent, OpenAI's Codex, and Google Antigravity at the same time through the built-in marketplace. Each agent gets direct access to IDE tools including project structure, symbol resolution, the emulator, and Logcat.

Google also released version 2.0 of the Android Bench leaderboard. It evaluates language models on multi-day development projects, scored on functionality, time spent, and cost.

Other additions:

  • Change Review: a verification tool that analyzes pending changes, generates bug reports, and explains modifications step by step.
  • Migration assistant (Canary): automatically translates older XML layouts into Jetpack Compose.
  • Lightbuild (testing): a declarative build system based on a YAML schema that both humans and AI agents can read and edit.

Android CLI cuts token usage

The Android CLI is now an official agent plugin for Claude Code, Codex, and Google Antigravity. New commands let agents render Compose previews in the terminal and check UI changes without booting an emulator. According to Google's internal tests, agents finished UI tasks 77 percent faster and used up to three times fewer tokens with this feature. The CLI can also stream physical test devices from Google labs.

Google built the open-source LeakCanary library into the CLI and the Android Studio Profiler. Heap dumps are evaluated on the computer rather than the phone. In a Google A/B test using Anthropic's Claude Opus 5.5, the CLI helped find and fix memory leaks nearly twice as fast while cutting token usage by about 25 percent.

Other commands let agents capture system traces via the Android Performance Analyzer, pull Firebase Crashlytics crash reports, and check code optimization through the R8 compiler. Google also provides more than 30 Android Skills, which feed documentation to agents for tasks such as checking Google Play policies before submission, handling permission security issues, and adapting layouts for foldables and laptops branded as Googlebooks.

Gemini Nano 4 and MCP for in-app agents

Google is adding an alpha Agentic API powered by Gemini Nano 4 to ML Kit for on-device processing. It executes functions automatically, saves chat history, manages the orchestration loop, and provides safety filters.

The AppFunctions API launches in beta as a unified tool-calling layer, using the Model Context Protocol (MCP), so apps can share functions with local or system-wide agents. A new alpha Agent-to-UI (A2UI) interface for Jetpack Compose lets agents show interactive interfaces built from an app's own components instead of plain text.

A preview of the Connected Apps Platform lets developers link apps through Android MCP or cloud-hosted MCP servers and manage them from a single console. Google says more than 50 partner integrations, including Adobe, Peloton, and Experian, are already live in the Gemini app and will roll out to web search in the coming weeks.

XR and XREAL AURA

The Jetpack XR SDK, the XR emulators in Android Studio, and the Engine Hub for Unity, Godot, and Unreal are now stable. Jetpack Compose 1.13 and the AppState model target large-screen adaptation.

With XREAL, Google unveiled the XREAL AURA, which XR chief Shahram Izadi calls the first wired smart glasses running Android XR. Specifications:

  • Weight under 95 grams
  • Transparent, dimmable lenses with a 70-degree field of view
  • External puck with trackpad, battery, and compute, running Qualcomm's new Snapdragon Reality Elite chip
  • Up to five Android apps in view at once, or laptop mirroring over USB-C
  • Gemini sees both the real-world environment and the virtual screens

Price is $1,099. Orders are open in the US, Canada, and South Korea, with shipping set to begin in mid-November.

What this means

Android Studio now treats coding agents as interchangeable components through ACP, with Claude Agent, Codex, and Antigravity running side by side against the same IDE tools. Much of the tooling announced here targets agent cost and speed: Compose previews without an emulator and LeakCanary analysis on the host machine are both framed around fewer tokens. The efficiency figures (77 percent faster, up to three times fewer tokens, nearly twice as fast, about 25 percent fewer tokens) come from Google's own tests and have not been independently reproduced.

On the app side, the on-device Agentic API is alpha, A2UI is alpha, AppFunctions is beta, and the Connected Apps Platform is a preview, so production use is limited for now. The web search rollout of partner integrations is described only as "in the coming weeks."

Related Articles

product update

Google launches Connected Apps Platform to bring third-party apps into Gemini and Search

Google announced the Connected Apps Platform, a unified route for developers to bring their apps into Gemini and Search. More than 50 third-party Connected Apps already exist in the Gemini app, and a Calm integration is due in the coming weeks.

model release

Google's Gemini 4 Carbon ranks #1 on Artificial Analysis Intelligence Index at 59, opens first to cyber defenders

Google announced Gemini 4 Carbon, which it says ranks #1 on v5.0 of the Artificial Analysis Intelligence Index with a score of 59. Access begins this week with a subset of cyber defenders through the Fairwind Program, at a stated price of $2 per million input tokens and $10 per million output tokens.

product update

GitHub adds 4M CPU cores and 160 PB storage, makes Copilot Fusion model router generally available

GitHub is adding more than 4 million CPU cores and 160 PB of storage as AI agents drive a surge in commits. At GitHub Universe it also made Copilot Fusion, a routing engine that picks frontier models per task, generally available.

model release

Google launches Genie 4 world model with audio and two-minute interactive sessions, powered by Gemini 3.8 Flash

Google announced Genie 4, a world model for generating interactive AI experiences that run about two minutes, up from one minute for Genie 3. It adds audio generation, real-time natural language updates and Gemini 3.8 Flash for its simulations. It launches as a research preview for trusted testers.

Comments

Loading...