Understand how Vultr organization administrators turn the MCP Server on or off in the Vultr Console and use the Use MCP permission policy to control access.
Admin controls let organization administrators decide whether the Vultr Model Context Protocol (MCP) Server is available to the organization and which identities in the organization can use it. The Vultr MCP Server is in public preview and is read-only. Connected AI assistants can view the account's resources but cannot create, modify, or delete anything.
MCP access is governed at two levels: whether MCP is available to your organization at all, and which users, groups, and roles are permitted to connect.
An administrator turns the MCP Server on or off for the entire organization in the Vultr Console, under Manage Organization and the MCP Server tab, or with the Vultr API. When MCP is disabled, no identity in the organization can connect to the MCP server, regardless of individual permissions.
In the Vultr Console, both enabling and disabling require an explicit confirmation step, because the change affects every connected AI assistant in the organization.
Enabling MCP for the organization makes it available. It does not grant any identity the ability to connect. Access is governed by the Use MCP permission policy (mcp.use), which you attach to the users, groups, or roles that can authenticate with the MCP server.
Granting this policy allows an identity to connect and use the read-only tools, and nothing more.
Use the following guides to create a policy and attach it:
0 Comments
Be the first to comment and share your perspective with the community.