Google unveils Gemini Spark to rival OpenClaw
At Google I/O 2026, the company introduced Gemini Spark, an always-on AI agent designed to operate in the background 24/7. Powered by the new Gemini 3.5 Flash model, the platform aims to automate personal tasks ranging from drafting correspondence to monitoring financial statements through persistent cloud-based virtual machines.

The system integrates directly with Google Workspace, including Gmail, Docs, and Sheets. To extend its reach, Google is adopting the Model Context Protocol, allowing the agent to interface with third-party services such as Canva, OpenTable, and Instacart. By this summer, the capability will expand to macOS, enabling the agent to interact with local files regardless of whether a device is active.
Josh Woodward, vice president of Google Labs, described the user experience as a continuous workflow where the agent catches tasks delegated by the user. Beyond standard app integration, Google plans to enable direct communication with Spark via text and email. Future updates will incorporate Chrome connectivity and a new interface dubbed Android Halo. While the agent functions autonomously, the company maintains that users retain granular control over which systems the AI accesses and when it remains active.
Comments (0)
No comments yet. Be the first!