What Is MCP (Model Context Protocol)?

    MCP is an open protocol that lets an AI model discover and call the tools of a software system — the difference between AI that talks and AI that acts.

    What is MCP?

    MCP, the Model Context Protocol, is an open standard that lets an AI model discover and call the tools a software system exposes. Instead of a developer wiring each capability by hand, the system advertises what it can do, and the AI selects and executes the right action within permissions.

    Key takeaways

    • MCP is a protocol for AI-to-software action, not a product.
    • The server advertises tools; the AI discovers and calls them.
    • It turns capabilities into something a model can reason about.
    • Permissions, confirmation and audit remain the server's job.

    What problem does MCP solve?

    Before MCP, connecting an AI assistant to a business system meant custom integration work for every capability, in every assistant, one at a time. Each new action was another development project.

    MCP inverts that. The business system runs an MCP server that publishes a list of tools — each with a name, a description and a typed input. Any MCP-capable AI can read that list and use it. Add a tool on the server and every connected assistant can use it immediately.

    How does an MCP call work?

    • The AI connects to the MCP server and requests the tool list.
    • The server returns tools with descriptions and input schemas.
    • The AI chooses a tool and supplies validated arguments.
    • The server authenticates the caller and enforces permissions.
    • The server executes, writes an audit record, and returns a result.

    Why does this change what AI can be?

    It moves AI from producing descriptions of work to performing work. The same request that previously ended with 'here is a draft you can copy' can end with a contract created, sent for signature and logged.

    The intelligence is not new. The permission to act is.

    Where URBLD fits

    URBLD runs an MCP server so an approved assistant can operate the platform through the same guarded actions the interface uses.

    Frequently Asked Questions

    More in MCP

    The protocol that lets AI do work instead of describing it.

    Browse MCP
    Share this page