Skip to main content

What MCP (Model Context Protocol) is and why it matters for property management software

MCP, the Model Context Protocol, is an open standard that lets AI apps such as Claude and ChatGPT connect to outside systems, including property management software. For an owner, it means you can ask your AI app about your real units, tenants and balances, and the software decides exactly what the AI may read or do. It matters because it turns a chat from guessing into answering from your records.

MCP in one paragraph

The official site, modelcontextprotocol.io, describes MCP as an open-source standard for connecting AI applications to external systems, and compares it to a USB-C port: one standard plug instead of a different cable for every device. Anthropic announced it as open source in November 2024, and the MCP site lists Claude and ChatGPT among the AI apps that support it.

Before a standard like this, every software company that wanted to work with an AI app had to build a custom connection for each one. With MCP, a property software company builds one MCP server, and any AI app that speaks MCP can use it.

The three parts: host, client and server

The MCP architecture page names three participants.

PartWhat it isIn property terms
HostThe AI application that coordinates connectionsClaude or ChatGPT on your phone or computer
ClientA piece inside the host that keeps one connection to one serverThe link between your AI app and one software account
ServerA program that provides context and actionsYour property software's connector, which knows your units and tenants

A server can run on your own computer or on the internet. Property software connectors are remote servers, reached over the web, and the MCP site says it recommends OAuth for obtaining the access tokens that remote connections use. OAuth is the familiar "sign in and press Allow" screen.

Tools, resources and prompts

An MCP server can offer three kinds of things. Tools are actions the AI app can call, such as "list vacant units". Resources are data the app can read, such as a document. Prompts are reusable templates. The AI app asks the server which tools exist, and the model picks one when your question needs it.

Example a made-up owner with 48 units asks, "How many units are vacant at Oak Street?" The AI app calls the software's find-units tool with the property and the status vacant. The server returns 3 units. The answer, 3 of 48, comes from the records, not from the model's memory, and the owner can open the software to see the same 3 units.

Why it matters for property management

  • Answers from records. Rent rolls, balances and vacancies are exact numbers. A connected AI app reads them instead of guessing.
  • No copy and paste of tenant data. You stop pasting names and balances into a chat by hand.
  • Permissions live in the software. The server decides what the AI can do. A careful vendor lets the AI read freely but keeps money and legal steps behind your own approval.
  • Your choice of AI app. One standard connector works with several AI apps, so you are not tied to one.
  • Less screen time. Many routine checks become one question on your phone.

What to ask a vendor about its MCP connector

  1. Which tools does it offer, and which of them only read?
  2. Can the AI app move money, sign anything or send a legal notice by itself? (The safe answer is no.)
  3. How do I sign in and choose which portfolio the app can see?
  4. How long does a connection last, and how do I disconnect it?
  5. Is every action the AI app takes written to a log I can see?

Ask for the list of tools in writing. A vendor that cannot list them clearly is not ready for you to trust with your portfolio.

What MCP does not do

MCP is a way to connect, not a promise about quality or safety. The MCP site notes that the protocol focuses only on how context is exchanged and does not dictate how AI applications use the model or manage what they receive. In practice that means three things for an owner. The AI app can still word an answer badly, so numbers that matter should be checked in the software itself. The connector is only as careful as the vendor who built it. And an AI app with no connector to your records is still guessing, however good the model is.

How DoorKeeping uses MCP

DoorKeeping offers an MCP connector for Claude and ChatGPT; setup takes one address and is described on Connect your AI app. Its tools cover the portfolio, leases, rent, repairs, messages, bills, banks and cards, and the books; the full list is on that page. Tools that only read answer right away, and anything that saves or sends shows you a draft first. Your AI app can never accept an approval: payments, signatures and legal notices wait for you to tap Accept on doorkeeping.com. You choose the portfolio on the Allow screen, a connection lasts at most 90 days, and every action is logged, as the security page explains.

Sources

  1. Model Context Protocol, What is the Model Context Protocol (MCP)?, read Oct 29, 2026
  2. Model Context Protocol, Architecture overview, read Oct 29, 2026
  3. Anthropic, Introducing the Model Context Protocol, read Oct 29, 2026

Questions

What does MCP stand for?
Model Context Protocol, an open standard for connecting AI apps such as Claude and ChatGPT to outside systems like property management software.
Is MCP safe for my property data?
MCP itself is a way to connect. Safety depends on what the vendor's server allows, so ask which tools it offers, whether it can move money, and how sign-in and logging work.
Do I need to be technical to use an MCP connector?
No. In most AI apps you add a connector by pasting an address, signing in and pressing Allow.