The ChMeetings MCP Connector allows you to access your ChMeetings data, including to create reports and gain insights, by asking questions in plain language to an integrated AI. 

The feature works with well known models such as ChatGPT, Claude, and Gemini; but also with any other AI tool that supports an MCP connection.
 

In this guide you will learn about:

• How This Works
• MCP Endpoints
• Read-Only Access to Data
• How to Connect
  • ChatGPT
  • Claude
  • More Tools
• Data Security with MCP
• Data Privacy when Using AI
• Data Accuracy when Using AI

How This Works

MCP (Model Context Protocol) is a technical connection standard that enables AI assistants to directly connect to in-app data and answer questions based upon it. 

For example, you can ask the AI who are the top three donors this year, how much did they donate, and to which funds? The AI will access your contribution and funds data, find the top 3 donors, and retrieve their donated amounts and corresponding funds. The result will be displayed to you within a chat conversation.
 

MCP Endpoints

Tools: http://mcp.chmeetings.com/tools 

Resources: https://mcp.chmeetings.com/resources 

The Resources MCP is intended for development tools such as VS Code or other AI coding assistants, when building applications that integrate directly with ChMeetings. Learn more.

Read-Only Access to Data

MCP usually allows read/write access to data. However, our connector strictly provides read-only access. The areas you can access data from are::

• People
• Families
• Groups
• Notes
• Giving
• Pledges
• Events
• Attendance

How to Connect

The connection process may differ depending on the tool of your choice. Below are some detailed guides for connecting to ChatGPT and Claude. We will also be linking to instructions on connecting MCP to Claude Code, Gemini CLI, and VS Code. If you are not using any of these tools, we recommend looking for their instructions on how to set up an MCP connection. 
 

Back to Contents

ChatGPT

1. Log into your ChatGPT account
2. Click Settings
   
    
3. Click on Plugins and click on Browse Plugins then click +
   
    
4. Enter the Name (ChMeetings) and Connection (http://mcp.chmeetings.com/tools) then check the “I Understand and Want to Continue” box - Remember that ChatGPT has no access to change your ChMeetings data via MCP. Click Create when done.
   
    
5. Sign into ChMeetings
   
    
6. Enter your API Key. See Enable the API to learn more about retrieving your key.
   
    
7. Here is example of a Question and Answer from ChatGPT:

Back to Contents

Claude

1. Click on Account and choose Settings
   
    
2. Click on Connectors then Add Custom Connector
   
    
3. Enter the title (which will be Chmeetings) then the URL (http://mcp.chmeetings.com/tools), and click Add
   
    
4. Click Connect. This will open a browser tab where you will be asked for your API Key. See Enable the API to learn more about retrieving your key.
   
    
5. Enter the API Key
   
    
6. Here is an example of a Question and Answer from Claude:
   

More Tools

• Claude Code
• Gemini CLI
• VS Code (GitHub Copilot)

Back to Contents

Data Security with MCP

Here is how ChMeetings ensures data security during an MCP connection:

• MCP is an open protocol that allows AI tools to interact with APIs more naturally. Instead of calling a specific ChMeetings API endpoint yourself, you simply ask a question in plain language. The AI translates your request into the appropriate API calls and returns the results. This is why an API key is required.
• The API key provides access to the entire ChMeetings account, so only Account Owners and Administrators can generate and use it. Other users do not have access to API keys, therefore cannot use the MCP integration and cannot access the integrated AI.
• MCP natively supports read/write access, but our current implementation strictly allows read-only access. This means your integrated AI cannot create, update, or delete any data in your ChMeetings account.
• The AI tool communicates with our HTTPS APIs, so all data is transmitted over encrypted connections, just like any other secure API request. The process is simple: AI Tool > ChMeetings API > Data Retrieved > AI Analysis > Response.
• ChMeetings does not interact with your integrated AI to generate any data you may retrieve via MCP. The AI service you choose retrieves the data itself from ChMeetings, through the MCP connection, and performs the analysis on its own. All ChMeetings does is provide read-only access to your data with no further intervention.
• If you've been exporting data from ChMeetings to Excel to then upload it to an AI tool for analysis, the MCP integration removes those manual steps by allowing the AI to securely access the data directly, making the process much faster and easier. From a security perspective, there is no fundamental difference between the two methods.
• ChMeetings does not require or recommend any specific AI tool and does not ask you to install additional software. You can use whichever AI platform they already trust and use regularly, as long as it connects to external apps via MCP.
• ChMeetings does not process or store your conversations with integrated AI or any  AI-generated outputs. Your AI provider does.
• ChMeetings does not sell or use customers' data by any means. This will never change.
   

Back to Contents

Data Privacy when Using AI

Connecting your ChMeetings data to an AI software means that your data must necessarily be passed to the company developing and managing the AI. This is because the AI cannot work as intended without accessing your data. This is why we strongly recommend that you seriously consider data privacy risks and mitigation measures specific to your organization before committing to connecting AI to your ChMeetings app.

The general default for any AI company is to use your data for model training and other purposes (some purposes may be stated clearly, others may be vague). Some companies give you a choice regarding what they may use the data for; while others do not.

If you choose to use their tools, we strongly recommend going through at least the following minimum due diligence before any live AI implementation: 

• Carefully check your AI app settings and adjust data use features that may be relevant to your organization. 
• Inform yourself about all possible scenarios under which your data may be used. 
• Explore AI settings in depth, and contact your provider for any technical clarifications. 
• Consult a technical professional to help you with any advanced know-how.
• Check company data policies to ensure you get the highest data privacy and security you can get under the circumstances. This means that you should:
  • Very carefully read the data policies of the AI company providing the tool you choose to implement. 
  • Contact your AI provider for any clarification of their data policies. 
  • Consult a legal professional before deciding to hand over your data to any AI.

Remember that your people’s data is very important. It may be highly private, sensitive and may expose vulnerabilities. We assume that your community, like most church communities, expects accountability, respect, and care; of which data protection is an integral part.
 

Back to Contents

Data Accuracy when Using AI

Please always keep in mind that AI may make mistakes, mismatch, distort, or outright fabricate data. This can range from people’s names to profile information, to attendance, giving, reporting, accounting, etc. All types of information are susceptible to being altered in AI outputs.

To clarify, here we specifically refer to the AI’s output/answer, which may alter data in the output/answer itself, not within the ChMeetings platform, where our implementation of MCP Tools does not allow write access. ChMeetings data cannot be changed via MCP, so you will always find your most accurate data, as entered by your staff, within the ChMeetings platform itself. We strongly recommend that you always double check all AI outputs against your actual ChMeetings data.

Back to Contents