Skip to content

MCP Server Pages in XTM One

This page complements MCP Servers in XTM One.

That guide explains what MCP servers are for. This page explains what the MCP server pages show and how to use them.

What you see in the sidebar

In the left navigation, this area appears as MCP Servers.

The list page title is MCP Tool Servers.

MCP servers list page

The list page is designed for quick operational scanning.

Use this page to:

  • search servers by name, command, or URL when you know the tool source but not how it was labeled in the workspace
  • filter by visibility to distinguish personal servers from shared or centrally managed ones
  • filter by tags to focus on one business capability, provider family, or operational context
  • compare transports so you understand whether the server runs as a subprocess or connects over HTTP
  • check whether a server is enabled before assuming its tools should be available to agents
  • see whether OAuth, headers, or environment setup is involved, which helps identify the likely maintenance path
  • add a new server when they have permission and need to expose a new external MCP toolset

The header includes:

  • page title MCP Tool Servers
  • short explanation that agents can use external tools through these servers
  • Add Server

Each card shows:

  • server name
  • transport label such as Subprocess (stdio), Streamable HTTP, or HTTP (SSE)
  • enabled or disabled state
  • company-managed or shared state
  • transport badge
  • command or URL preview
  • env badge when environment variables are present
  • OAuth badge when applicable
  • tags or argument count

This lets the list page answer two important questions immediately:

  • is the server likely usable right now
  • what kind of setup is behind it if something goes wrong later

Empty states on the list page

If no servers exist yet, the page invites the user to add a first MCP server.

If filters remove every visible server, the page shows No servers matching your filters.

MCP server detail header

When you open a server, the detail header shows:

  • back button to MCP Servers
  • server name
  • transport label
  • Enabled or Disabled
  • OAuth connection status when applicable
  • company-managed, shared, and read-only badges when relevant

The main actions in the header are:

  • Test Connection
  • delete

Test Connection is usually the first action to try after any setup or credential change, because it tells the user whether the issue is with the server connection itself or somewhere later in the agent workflow.

Test Connection is the main confidence-building action on this page.

Tabs on the detail page

The detail page contains:

  • Overview
  • Configuration
  • Activity

Overview tab

The Overview tab shows whether the server is meaningfully ready for use.

Its KPI cards show:

  • agents using the server
  • discovered tool count
  • tool call count
  • last activity

The overview card shows the connection details:

  • command and arguments for stdio
  • endpoint URL for remote transports
  • whether environment variables are set
  • whether custom headers are set
  • OAuth state when OAuth is used
  • tags
  • created and updated timestamps

The bottom of the tab contains an Agents Using card so you can see downstream usage immediately.

Configuration tab

The Configuration tab is the most operationally important screen.

The main configuration card covers:

  • server name
  • transport choice
  • command or URL
  • environment variables or HTTP headers
  • tags
  • Hide from chat
  • sharing and visibility settings
  • public catalog state when available

In edit mode, you switch transport directly in the UI between:

  • stdio
  • Streamable HTTP
  • SSE

The same screen also contains the Discovered Tools card.

That card lets you:

  • discover tools for the first time
  • refresh discovered tools
  • read each tool description
  • test an individual tool

This matters because many questions about MCP are not really about the server object itself. They are about whether the expected tool is actually visible and callable.

The tool card also shows useful operational states:

  • loading while connecting
  • error banner when discovery fails
  • empty state when no tools are discovered yet

OAuth connection card

When the MCP server uses OAuth, the Configuration tab also includes an OAuth Connection card.

That card shows:

  • whether the server is connected
  • whether token refresh failed
  • whether authorization is still required
  • authorize or re-authorize action when allowed

This card is where you can see whether the server is configured correctly but still not authenticated.

Activity tab

The Activity tab is the traceable history area for the server.

Users typically open it when they need to confirm:

  • whether the server changed recently
  • whether calls were made recently
  • whether a connection or tool issue started after an edit

Shared, company, and read-only states

MCP server pages are strongly influenced by ownership rules:

  • private servers are usually editable by their owner
  • shared servers are reused across group members
  • company-managed servers are visible more widely but read-only for non-admins

This matters because testing may be available while editing is not.

Good habits on these pages

  • Use Test Connection before troubleshooting the agent itself.
  • Refresh discovered tools after changing transport or credentials.
  • Read the discovered tool list, not just the server name.
  • Notice the difference between a configured server and an OAuth-connected server.