Google Gemini Mac App Adding Spark Agent and Voice Control This Summer

Google announced at I/O 2026 that its Gemini Mac app will receive two major features this summer: the Spark AI agent and enhanced voice control capabilities.

Spark Agent Coming to macOS

Spark is an AI agent that can take actions on behalf of users to "navigate your digital life," according to Google. On macOS, Spark will be able to perform tasks involving local files and automate workflows across the desktop. This extends the existing capability to use open windows as context for prompts.

The agent integrates with Gmail, Google Docs, and other Workspace apps, as well as third-party services. Spark requires a Google AI Ultra subscription, priced at $100 per month. The feature launches in beta next week on Android, iOS, and web, with the Mac version following this summer.

Enhanced Voice Control

The new voice experience allows users to speak naturally without worrying about filler words like "ums" or "what abouts." Users long-press the function key on Mac to activate Gemini, which displays a floating pill at the bottom of the screen. Releasing the key submits the prompt.

Using screen context, Gemini can convert free-flowing speech into precise drafts and automatically reformat text at the cursor position. Google demonstrated selecting files in Finder and dictating an email that automatically populates a Gmail compose window.

Background

Google launched the native Mac app in April, developed by a small team using the company's Antigravity framework. The app already supports using open windows as contextual input for prompts.

What This Means

Google is positioning Gemini as a comprehensive desktop AI assistant rather than just a chatbot. The $100/month Ultra pricing puts it at the premium end of AI subscription services. The local file access and desktop automation capabilities represent a significant expansion beyond web-based AI assistants, though privacy and security implications of granting an AI agent access to local files and desktop workflows remain to be seen. The voice control improvements address a common pain point in voice interfaces, but the feature's accuracy with natural speech patterns will determine its practical utility.