Google is launching managed MCP servers that allow AI agents to simply connect to its tools

AI agents are sold as a solution for planning trips, answering business questions, and solving problems of all kinds, but getting them to work with tools and data outside of their chat interfaces has been difficult. Developers have to patch different connectors and keep them running, but this is a brittle approach that is difficult to scale and causes governance issues.

Google claims that it is trying to solve this problem by launching remote, fully managed MCP servers that will make Google and cloud services – such as Maps and BigQuery – easier for agents to connect to.

The move comes on the heels of the launch of Google’s latest Gemini 3 model, and the company looks to combine stronger logic with more reliable connections to real-world tools and data.

“We’re making Google Agent ready by design,” Strain Giannini, director of product management at Google Cloud, told TechCrunch.

Instead of spending a week or two setting up connectors, developers can now essentially paste a URL to a managed endpoint, Giannini said.

At launch, Google starts with MCP servers for Maps, BigQuery, Compute Engine, and Kubernetes Engine. In practice, this might look like an analytics assistant querying BigQuery directly, or an operations agent interacting with infrastructure services.

In the case of maps, Giannini said, without MCP, developers would rely on the knowledge built into the model. “But by giving your agent […] “With a tool like the Google Maps MCP server, we rely on actual and updated location information for places or planning trips,” he added.

Although MCP servers will eventually be offered across all Google tools, they are initially running under public preview, meaning they are not yet fully covered by Google Cloud’s terms of service. However, it is offered to enterprise customers who already pay for Google services at no additional cost.

“We expect to make it available to the public very soon in the new year,” Giannini said, adding that he expects more MCP servers to appear every week.

MCP, which stands for Model Context Protocol, was developed by Anthropic about a year ago as an open source standard for connecting AI systems to data and tools. The protocol has been widely adopted throughout the world of agent tools, and earlier this week Anthropic donated MCP to a new Linux Foundation fund dedicated to open source and standardizing infrastructure for AI agents.

“The beauty of MCP is that, because it’s a standard, if Google provides a server, it can connect to any client,” Giannini said. “I look forward to seeing how many other customers show up.”

One can think of MCP clients as AI applications on the other end of the wire that talk to MCP servers and communicate with the tools they provide. For Google, this includes the Gemini CLI and AI Studio. Giannini said he also tried this with Anthropic’s Claude and OpenAI’s ChatGPT as clients, and “they just worked.”

Google says it’s not just about connecting agents to its services. The big game for enterprises is Apigee, the application programming interface (API) management product, which many companies already use to issue API keys, set quotas, and monitor traffic.

Apigee can essentially “translate” a standard API into an MCP server, turning endpoints like the product catalog API into tools that the agent can discover and use, layering existing security and governance controls on top, Giannini said.

In other words, the same API guardrails that companies use for human-created applications can now apply to AI agents as well.

Google’s new MCP servers are protected by a permission mechanism called Google Cloud IAM, which explicitly protects what a proxy can do with that server. It’s also protected by Google Cloud Model Armor, which Giannini describes as a dedicated firewall for proxy workloads that defends against advanced proxy threats like hotshot injection and data filtering. Administrators can also rely on audit logging for additional observability.

Google plans to expand MCP support beyond the initial set of servers. In the next few months, the company will provide support for services across areas such as storage, databases, logging, monitoring and security.

“We built in the plumbing so developers don’t have to do it,” Giannini said.