Google also previewed a remote task feature that is expected to arrive later. Once available, users will be able to assign multi-step tasks from their phones for Spark to execute on their Mac while they are away, such as locating a file, extracting information from it, and emailing the results.
The desktop expansion positions Gemini Spark alongside other AI desktop agents capable of interacting with local files and applications. According to Google’s announcement, Gemini Spark for macOS is available in beta to Google AI Ultra subscribers aged 18 and older in the United States.
Google is also expanding Spark’s reach across additional apps and services. The AI agent now integrates with Google Tasks and Google Keep, allowing users to turn notes stored in Keep into actionable items in Tasks. Support is also rolling out for Canva, Dropbox, Instacart, OpenTable, and Zillow Rentals, enabling Spark to complete tasks such as designing flyers, accessing and sharing files, reserving restaurant tables, ordering groceries, and scheduling apartment tours. Google said these integrations will begin rolling out on the web and mobile versions of Gemini Spark over the next week before arriving on macOS in the coming weeks.
Another addition is support for custom Model Context Protocol, or MCP, which lets users connect their own applications directly to Gemini Spark to create a more customized assistant.
Google is also giving Gemini Spark the ability to monitor topics continuously and respond to events as they happen. Users can ask Spark to follow sports, financial markets, blogs, news sites, social media, shopping, weather, and email, then receive updates when specific conditions are met. Google said Spark can, for example, deliver match analysis after a soccer game ends or send a financial report when a stock reaches a chosen threshold.
The latest updates continue Google’s push to expand Gemini Spark beyond a conversational assistant into an AI agent capable of carrying out tasks across desktop software, cloud services, and connected applications, while adding new automation features that reduce the need for users to manually monitor information or switch between multiple tools.
This analysis is based on reporting from Google.
Image courtesy of Google.
This article was generated with AI assistance and reviewed for accuracy and quality.
