Gemini Spark, Google's always-on personal AI agent, is picking up support for third-party apps, including the Model Context Protocol, along with a new real-time topic-tracking feature, expanding what the assistant can do beyond its existing macOS app support.
Gemini Spark's new third-party app connections
On the first-party side, Spark can now reach into Google Keep and Google Tasks, letting users, in Google's words, scan their scattered brain-dumps in Keep and turn them into action items in Tasks. That closes a gap for anyone who jots down quick notes throughout the day but rarely circles back to turn them into an organized to-do list.
The bigger expansion, though, is third-party support for Canva, Dropbox, Instacart, OpenTable, and Zillow Rentals. According to Google, that means users can design custom flyers, access and share files, reserve a table for date night, order weekly groceries, or reserve an apartment tour, all without leaving the Spark interface. The rollout starts on web and mobile over the next week, with Mac support arriving in the coming weeks.
MCP support opens the door to more integrations
Alongside the named app partners, Google is also rolling out Model Context Protocol support for Gemini Spark. MCP has become a common standard for connecting AI assistants to external tools and data sources, and Google's implementation lets users add a custom app link directly at the bottom of gemini.google.com/apps to get started.
That MCP support matters beyond the handful of apps Google named at launch. Rather than waiting for Google to build and ship a dedicated integration for every service, developers and power users can wire up their own tools to Spark as long as those tools expose an MCP-compatible interface. It's the same approach several competing AI assistants have adopted to avoid being limited to a fixed list of first-party partners.
Real-time topic tracking is Spark's other big update
Google is also giving Gemini Spark the ability to intelligently track topics and react to events in real time. In practice, that covers things like sending a recap the moment a favorite soccer match ends, or firing off a detailed report if a tracked stock crosses a set price threshold.
Google frames the feature broadly: Spark can be asked to keep an eye on blogs, news sites, social media, finance, shopping, weather, and sports, in addition to monitoring email, so users don't have to keep manually refreshing pages to catch updates. That positions Spark less as a chatbot that only responds when prompted and more as a background agent that surfaces information proactively.
Why this pushes Gemini Spark closer to a true personal agent
Between app integrations, MCP support, and real-time tracking, Google is clearly building toward an assistant that can act across a person's entire digital life, notes, tasks, files, reservations, groceries, and ongoing events, rather than one confined to a single chat window. That kind of always-on, cross-app agent is where much of the broader AI industry has been racing to compete, from dedicated shopping assistants to scheduling bots.
For now, Gemini Spark remains exclusive to Google AI Ultra subscribers, so the expanded app support and real-time tracking won't reach every Gemini user immediately. Anyone curious about trying it will need to be on Google's top-tier AI subscription and wait for the web and mobile rollout to reach their account over the coming week.
Quick FAQ
Which apps does Gemini Spark now support?
Spark now works with Google Keep and Tasks as first-party integrations, plus third-party support for Canva, Dropbox, Instacart, OpenTable, and Zillow Rentals.
Do I need a special plan to use Gemini Spark?
Yes, Gemini Spark is currently available only to Google AI Ultra subscribers.
