What Gemini Spark on macOS is
Gemini Spark (macOS): overview
Gemini Spark is an AI agent built into Google's Gemini that does work for you. Now that Spark has come to the Mac Gemini app, you can hand off tedious tasks on your desktop. For how AI operates real environments, our guide to Claude controlling robots (Project Fetch) shows the wider reach of agents.
An agent that automates desktop tasks
The point of Mac support is to take over the manual busywork that piles up on a PC. Google brought Spark to the Mac Gemini app so it can automate time-consuming tasks across your desktop.
"We're bringing Spark to the Gemini macOS app to help you automate time-consuming tasks across your desktop." — from the Gemini Spark macOS announcement
For example, sorting a pile of scattered PDFs into folders or building a budget spreadsheet from multiple invoices are cited as the kind of routine, time-consuming tasks it targets.
Eligibility: US AI Ultra subscribers (18+)
Availability is limited at first. The Mac version of Spark is available in beta to US Google AI Ultra subscribers aged 18 and over.
"Gemini Spark for macOS is available in Beta to Google AI Ultra subscribers aged 18 and over, starting in the US." — from the eligibility statement
It's positioned for the higher-tier AI Ultra plan and US-first, starting with a narrow audience.
New connected apps and MCP support
Newly added connections
Spark's value depends on how many apps it connects to so it can work across them. This update widened those connections.
Canva, Dropbox, Instacart, OpenTable, Zillow Rentals
First, integrations with external apps were added. New connections were added with Canva, Dropbox, Instacart, OpenTable, and Zillow Rentals, along with Google Tasks and Keep.
"We're also launching integrations with Canva, Dropbox, Instacart, OpenTable and Zillow Rentals" — from the connected apps section
Spanning design, file storage, shopping, reservations, and home search, these apps cut across life and work, making it easier for Spark to carry a task across multiple services.
Custom MCP support to connect your own apps
Beyond the ready-made connections, there's now a path to connect your own. Google is rolling out custom MCP support, letting you connect your favorite apps directly into Spark.
"We're also rolling out support for custom Model Context Protocol (MCP), giving you the ability to connect your favorite apps directly into Spark" — from the MCP support section
MCP is a common standard for connecting AI tools to external apps. Supporting it means even apps Google hasn't pre-integrated can become work targets for Spark, as long as they follow the standard.
Rollout timeline and features
Connected-apps rollout schedule
Finally, here is when it reaches you and what new features come with it.
Rollout schedule (a week for web/mobile, weeks for macOS)
The connected apps arrive in stages. They roll out to Spark on web and mobile over the next week, and to the macOS app over the coming weeks.
"These connected apps are rolling out over the next week on Gemini Spark on web and mobile and will be rolling out to the macOS app in the coming weeks." — from the rollout timeline
The Mac version takes a little longer, but it reaches eligible users in turn.
Real-time tracking and what's next
On the feature side, Spark also gained the ability to keep up with information. Gemini Spark can now intelligently track topics and react to events in real time.
"We're also giving Gemini Spark the ability to intelligently track topics and react to events in real time." — from the real-time tracking section
In short, this update expands Gemini Spark into an agent that "automates Mac desktop tasks, connects to more apps, and keeps up with changing information." The same June 30 announcements also included the image and video model in our guide to Nano Banana 2 Lite, showing Google advancing AI on both the app and base-model fronts.
When gathering material or references for Spark from the web, converting pages to Markdown first makes them easier for the AI to handle and improves accuracy.
