Gemini Spark gets macOS launch and connected apps for agentic tasks

Google expanded Gemini Spark with a native macOS launch, connected apps, and additional updates. The headline capability is that Spark can now handle ongoing agentic tasks, with Google citing an example of monitoring interior design internships — a persistent, background job rather than a one-shot query. This pushes Gemini's assistant beyond conversational Q&A into proactive, long-running desktop workflows.

The macOS launch is strategically notable: bringing Gemini natively to Apple's desktop platform expands Google's reach into a user base where it competes directly with Apple Intelligence and any future Siri-Gemini integration. Connected apps let Spark act across a user's tools, deepening its utility as a genuine agent rather than a chatbot.

The move fits the industry-wide agentic pivot on display across this briefing — AWS WorkSpaces for agents, Anthropic's agentic Sonnet 5, and OpenAI's Codex hardware all point to the same destination: AI that acts, not just answers. Google's advantage is its data graph and app ecosystem, which it is leveraging aggressively (see also free personalized image generation).

As a desktop agent, Spark's value will depend on reliability and the breadth of connected apps, and on whether users trust it with the access required to act on their behalf. The macOS expansion is incremental relative to Google's bigger Gemini news this week, but it's a meaningful step in seeding Gemini across every surface. Watch for how deeply Spark integrates with third-party apps and whether Google extends it beyond macOS.