Skip to content

Developers · 6 min read

What is an MCP server? A practical guide

The Model Context Protocol (MCP) is an open standard for connecting AI applications to tools and data. An MCP server exposes a set of tools; an MCP client, such as an AI assistant or agent, discovers and calls them.

7-day trial with a card on file and 300 usage credits. Then $49.99/month unless you cancel.

How MCP works

An MCP client connects to a server, lists the tools it offers, and calls them with structured arguments. Modern servers use Streamable HTTP, so they can run as ordinary web services behind authentication.

Why scoping matters

An MCP server is a doorway into your systems. Each token should carry only the scopes it needs, act for a specific person, and stop working when that person's access is removed. Financial approvals and other sensitive actions should stay in the authenticated app, not in a tool call.

  • Read scopes for reporting and analysis.
  • Agent scopes for creating runs and schedules.
  • Publishing scopes only where reports must be released.
  • Revocation that takes effect immediately.

Common uses for agencies

Let a coding assistant or another agent read client reports, start a research run, check credit balances or draft a report, all within the permissions of the person who created the token.

Versabot's MCP server

Versabot exposes 21 agency tools over Streamable HTTP with scoped, hashed tokens, current-role enforcement and revocation, verified with the official MCP SDK.

Frequently asked questions

Is MCP only for Claude?

No. MCP is an open protocol supported by a growing number of AI clients and frameworks.

Published by the Versabot team.

Try an accountable AI agent on real client work.

Connect what you already use, ask for the work in plain language, and approve what leaves the building.

7-day trial with a card on file and 300 usage credits. Then $49.99/month unless you cancel.