Updated 3 minutes ago

Share this article

Related News

Apigee Hybrid brings MCP tool traffic inside customer Kubernetes clusters

Oct 14, 2026

Apigee Hybrid brings MCP tool traffic inside customer Kubernetes clusters

Apigee Hybrid 1.17.0 can run its MCP data plane inside a customer Kubernetes cluster, keeping runtime tool calls local while Google’s control plane supplies configuration.

Google CloudApigee HybridModel Context Protocol

Google unveils Gemini work agent, with key features still in early access

Oct 8, 2026

Google unveils Gemini work agent, with key features still in early access

Google’s new work agent promises delegated tasks across business software. Its availability notes and access-control documentation reveal what teams should verify before relying on it.

GoogleGemini EnterpriseAI agents

Meta’s WhatsApp Business MCP: setup, test messages and production limits

Sep 16, 2026

Meta’s WhatsApp Business MCP: setup, test messages and production limits

Meta’s September 15 MCP launch connects coding agents to WhatsApp setup and test messages. The tool references show where admin permissions, webhook tests and a separate Cloud API integration divide the work.

WhatsApp BusinessModel Context ProtocolMeta

Gemini Enterprise can render MCP Apps as interactive interfaces inside chat — Source: OpenTools Team / original MCP Apps connect chat, interface and server evidence mapSource image

Interactive Agents

Gemini Enterprise can render MCP Apps as interactive interfaces inside chat

By OpenTools Team

Google’s new developer guide shows how MCP Apps pair a tool with a live interface rendered in Gemini Enterprise. The pattern can shorten multi‑step work, but developers must still bind the widget to approved tools, isolate it in a sandbox and verify every consequential server action.

MCP Apps add a visual surface to an agent tool

Google’s October 19 developer guide describes MCP Apps as an extension that pairs a Model Context Protocol tool with a live interface. In Gemini Enterprise, the interface can render inside the conversation in a sandboxed iframe and expand into a side panel or fullscreen view.

The value is clearest for work with visual state or many small adjustments. A user can change filters, drag controls or inspect a preview locally instead of describing every adjustment in another prompt. The assistant and interface remain synchronized around the same task.

This does not turn arbitrary HTML into a trusted action surface. The interface is a client for named tools, and the server still decides what those tools are allowed to read or change.

The app has three connected pieces

Google’s implementation path starts with an MCP server that exposes a POST /mcp endpoint. The server declares tools and also publishes a ui:// resource containing the bundled widget. Tool metadata links a tool result to that resource so the client knows what interface to render.

Google suggests hosting the server on a private Cloud Run service while Gemini Enterprise supplies workforce sign‑in and governed access. The widget runs in the client, but consequential work should travel through the MCP server rather than through hidden browser credentials or direct database calls.

Keep the tool schema narrow. If the interface edits a project schedule, expose explicit read and update operations with stable identifiers and expected revisions rather than a generic command field that can reach unrelated records.

Local interactivity should not weaken server controls

A responsive widget can validate a date, preview a chart or maintain unsaved form state without an agent round trip. The server must repeat all security‑relevant validation: identity, tenant, object ownership, field limits, current revision and authorization for the exact action.

Treat values arriving from the widget as untrusted input even when the iframe is sandboxed. Sandboxing limits what the embedded interface can do to the surrounding client; it does not prove that a submitted value is correct or authorized.

For writes, return the new saved revision and render it back from a fresh server read. Use idempotency keys for repeat submissions and show an explicit conflict when the record changed since the interface loaded.

Test the conversation, widget and server as one flow

A complete acceptance test starts in Gemini Enterprise, invokes the tool, confirms the correct widget appears and checks that resizing or fullscreen does not hide essential labels. Then exercise empty, long, invalid and stale values, including a second submission of the same change.

Verify that screen‑reader names, keyboard focus, color contrast and mobile sizing survive the embedded context. Inspect server logs for the principal and tool name without recording secrets or unnecessary user content. Finally, read the changed record through its supported API and prove that an identical repeat is a no‑op.

MCP Apps can make agent work feel more like a focused application. Their reliability depends on treating the interface as one layer of a controlled system, with the authoritative checks at the server boundary.

Figure by OpenTools Team; no third‑party expressive material used. It maps the MCP tool, UI resource and hosted server to their separate responsibilities.

Sources

1.cloud.google.com (cloud.google.com)
2.modelcontextprotocol.io (modelcontextprotocol.io)

Tags

Gemini EnterpriseMCP AppsModel Context ProtocolCloud RunAI development