Google DeepMind brings MCP support to Gemini models and SDK

Connecting AI models to external tools shouldn't require reinventing the wheel every single time.

Google DeepMind CEO Demis Hassabis announced on X that Google will officially support the Model Context Protocol (MCP) across its Gemini models and SDK.

Listening to the community

Hassabis praised MCP as a solid protocol that is quickly emerging as the open standard for the agentic AI era. He added that Google plans to collaborate directly with the MCP team and the broader industry to help build it out further.

The move follows vocal requests from developers. Google CEO Sundar Pichai previously responded to community calls online with enthusiasm, confirming the decision to adopt MCP after listening to feedback.

Why it matters

We think this is a smart, pragmatic play from Google that puts developers first.

For builders, this eliminates a ton of unnecessary friction. Instead of writing bespoke integrations just to let Gemini talk to internal databases, local files, or specific APIs, developers can now rely on a single, standardized interface across multiple platforms.

For the AI ecosystem, it is a huge milestone. MCP was originally spearheaded by Anthropic, and Google adopting it gives the protocol massive legitimacy. When key frontier labs coalesce around one shared open standard, it prevents the agent ecosystem from fracturing and lets everyone focus on building better agents rather than proprietary plumbing.