Ever wish the AI assistant Claude knew your stuff? Like, your work documents, your project notes, or your study guides? Instead of just knowing about the whole wide world, it could know about your world.
Refer to our Claude Desktop + CustomGPT.ai’s Hosted MCP server docs for more info.
Well, now it can. By connecting CustomGPT.ai to your Claude Desktop app, you can give it a private brain loaded with your own information.
What’s this “MCP” thing anyway?
Think of the Model Context Protocol (MCP) like a USB-C cable for AI. It’s a standard plug that lets any AI app (like Claude) connect to any data source (like your CustomGPT.ai project) without a bunch of custom wiring. One plug, endless possibilities.
This means you can easily and securely let Claude peek into your private files to find answers.
Why is CustomGPT.ai the perfect “brain” for Claude?
- It’s accurate: We’ve been ranked #1 for finding the right answers in business documents. Less nonsense, more facts.
- It’s hands-off: We handle all the techy server stuff, so you don’t have to.
- It’s free with your plan: This powerful feature doesn’t cost you an extra dime.
Want to get into the nitty-gritty? Check out our Hosted MCP launch blog post.
How the Magic Happens
Here’s a simple map of how your question gets an answer from your files:
You, asking a question in the Claude app
│
▼
┌─────────────────────────────────┐
│ Your PC runs a little helper │
│ (called `supergateway`) │
└─────────────────────────────────┘
│ (Sends question securely)
▼
┌─────────────────────────────────┐
│ CustomGPT.ai Server │
│ (Searches your private files) │
└─────────────────────────────────┘
│ (Sends answer back)
▼
┌─────────────────────────────────┐
│ Your PC's helper gets the │
│ answer │
└─────────────────────────────────┘
│
▼
The answer appears in your Claude chat!
What you’ll need first:
- A CustomGPT.ai account (a free trial works fine!).
- A project in CustomGPT.ai with your files uploaded (PDFs, docs, etc.).
- The Claude Desktop app (version 1.3.0 or newer). This won’t work on the Claude website.
- Node.js (version 18 or higher) installed on your computer.
Let’s Hook It Up: Step-by-Step
- Get your secret token: Log in to CustomGPT.ai, open your project, and go to Deploy ➞ MCP Server (Beta). Click Generate MCP Token.
- Open Claude’s config: In the Claude Desktop app, click the three dots (⋯) for Settings ➞ Developer ➞ “Edit Config”.
- Paste the code: Paste this block into the file. Make sure to replace <PROJECT_ID> and <TOKEN> with your details from the CustomGPT.ai dashboard.
{
"mcpServers": {
"customgpt-mcp-server": {
"command": "npx",
"args": [
"-y",
"supergateway",
"--sse",
"https://mcp.customgpt.ai/projects/<PROJECT_ID>/sse",
"--header",
"Authorization: Bearer <TOKEN>"
]
}
}
}- Save and Restart Claude.
- Give it permission: When you first use it, Claude will show a banner asking to “Allow external integration?” Click Allow for this chat.
That’s it! Now, to get an answer from your own data, just ask a question and add “…answer from your knowledgebase” at the end.
Watch it in action:
Your Turn to Build!
So there you have it. No more generic answers! Your favorite AI tool is now your personal expert, with a direct line to the documents and data you care about most. It’s a true game-changer for getting things done.
Why not give it a try? Head over to CustomGPT.ai to start a free trial and create your own private AI brain.
We can’t wait to see what you build with it!
P.S If are into MCP world, you will definitely live this MCP AMA that we did with Pinecone folks and other MCP folks (Santiago).
Find it here:
Feel free to subscribe our Youtube Channel for more MCP related informative content.
Further Reading
For those interested in exploring MCP in greater depth, check out the following resources:
- What Is Model Context Protocol (MCP)?
A comprehensive introduction to MCP’s origins, goals, and why it’s becoming the de facto “USB-C” for AI integrations. - Top 124 MCP Servers, MCP Clients & Tools (2025 Guide)
A curated catalog of production-ready MCP servers and clients—both open-source and proprietary—showcasing plug-and-play tools you can use today. - Problems MCP Model Context Protocol Solves
An exploration of the real-world pain points (API sprawl, ad-hoc integrations, security gaps) that MCP was designed to eliminate. - The Model Context Protocol (MCP) Architecture (2025)
A deep dive into MCP’s component layers, sequence diagrams, security considerations, and production-grade deployment patterns.
Frequently Asked Questions
If I ask questions in Claude Desktop, will it use my CustomGPT.ai knowledge?
Yes. After you connect Claude Desktop to CustomGPT.ai through MCP, Claude can use your project’s private information (such as work docs, notes, or study materials) as context for answers instead of relying only on general world knowledge.
What do MCP permissions mean when connecting Claude Desktop?
Permissions determine what data Claude can access through the MCP connection. In practice, you should only connect approved knowledge sources so Claude can securely access relevant private files.
Is MCP mainly for querying existing knowledge, or for uploading files too?
In this setup, MCP is presented as the connector between Claude and an existing CustomGPT.ai data source (your project). A reliable workflow is to keep your source content in the CustomGPT.ai project, then query it from Claude via MCP.
Why use a hosted MCP server for Claude Desktop instead of custom wiring?
A hosted MCP setup reduces integration effort because MCP is a standard connection layer, and server operations are managed for you. This is useful if you want faster setup and less infrastructure work.
How can you improve answer quality after connecting Claude to CustomGPT.ai?
Start by loading Claude with the right private knowledge domain (for example, your work documents, project notes, or study guides) through your CustomGPT.ai project. Quality improves when the connected source matches the questions you expect to ask.
Is connecting Claude Desktop to a hosted MCP server suitable for business use?
It can be suitable when you need Claude to access private organizational knowledge in a secure way. The integration is explicitly described as secure access to private files, and it is positioned for business-document accuracy and enterprise use cases.
Priyansh is Developer Relations Advocate who loves technology, writer about them, creates deeply researched content about them.