Imagine you're deep in coding mode, trying to fix a bug, and you need to reference your project scope—except it's buried in a Google Doc. Or you’re about to implement a new feature, but the Jira ticket lacks key details, which are in a Slack discussion from two days ago. Instead of switching tabs and searching manually, what if Cursor or Windsurf could automatically pull that context into your IDE?

MCP (Model Context Protocol) makes this possible. MCP is a universal interface that allows AI assistants to call external data sources and APIs effortlessly.

In this tutorial, you’ll learn what MCP is, why it’s useful, and how to seamlessly connect private data from Google Drive, Jira, Slack, and more to your IDE for easy access and a much improved development workflow!

Before getting started, be sure to create a free Ragie account (this is what we’ll be using to connect your confidential data sources into your IDE workflow).

What is MCP and Why Does It Matter?

Think of MCP as a universal power strip for AI applications. Traditional APIs are like individual power adapters—each one designed for a specific device. MCP, on the other hand, provides a single connection point that allows AI models to access multiple data sources and tools without needing a custom adapter for each. While traditional APIs (OpenAPI, REST, GraphQL) already exist, they often require extensive manual integration. MCP simplifies this by offering a standard format that AI tools can understand and interact with. 

But Why MCP Instead of OpenAPI, REST, or GraphQL?

Traditional API specs like OpenAPI and GraphQL are designed for structured programmatic access, but AI agents often work differently:

• They don’t always know which API to call in advance.
• They need a generalized way to query different tools without hardcoding specific APIs.
• They often operate on natural language descriptions instead of rigid API endpoints.

MCP bridges this gap by offering a way for AI agents to dynamically discover and interact with external systems without developers needing to write glue code for every integration.

How to Hook External Services Into Your Cursor IDE

Going back to our previous example, let’s say you’re a software engineer working on a Jira ticket that references a project requirement document stored in Google Drive. Typically, you'd have to:

1. Open Jira and read the ticket details.
2. Manually search Google Drive for the referenced document.
3. Open Slack to find any related discussions.
4. Then actually start working on the code.

With Ragie’s MCP server, your AI assistant in Cursor can handle all of that for you:

• Find the project specification in Google Drive.
• Understand your internal APIs from project documentation.
• Surface past bug reports, customer requests, and design notes.
• Summarize relevant Slack messages.
• Pull in related Jira tickets for context.
• Then ship a contextually relevant code. 

All within your IDE. Ragie is a RAG-as-a-service platform that makes it easy to connect AI models to your knowledge sources. Ragie’s MCP server acts as a bridge between AI tools like Cursor and external services like Jira, Google Drive, and Slack—allowing them to retrieve and interact with information seamlessly. 

‍Step 1: Connect Your Data Source To Ragie‍

If you didn’t already, create a free account on Ragie using this link: https://secure.ragie.ai/sign-up 
‍
Then, follow the example below to connect your data source and files via Ragie Dashboard:

• Google Drive for specs and design docs
• Slack for team discussions and clarifications
• Jira for tickets and issue tracking
• Confluence, Notion, and others for internal documentation

Note: When the import mode is set to hi_res, images and tables will be extracted from the document. This is useful if you need your AI coding assistant to be able to analyze images and tables within your documents. The fast mode only extracts text and can be up to 20 times faster than hi_res.

After connecting your data sources to Ragie’s RAG pipeline, ensure you have all the prerequisites below before you continue to follow the tutorial. 

Step 2: Connect Ragie's MCP to Cursor

Once you’ve connected your data sources in Ragie (Step 1), there are two ways to connect Ragie’s MCP Server to Cursor. We recommend starting with Option 1 for simplicity, but if Cursor throws an error (known issue), you can fall back to Option 2 (using a shell script).

Option 1: Create an mcp.json File

You can add this configuration in either of these locations depending on where you want to use Ragie MCP Server:

• Project-specific: .cursor/mcp.json (in your project directory), for tools only available in that project.
• Global: ~/.cursor/mcp.json (in your home directory), for tools available in every Cursor project. 

Example mcp.json

{
  "mcpServers": {
    "ragie": {
      "command": "npx",
      "args": [
        "-y",
        "@ragieai/mcp-server"
      ],
      "env": {
        "RAGIE_API_KEY": "your_api_key",
        "RAGIE_PARTITION": "optional_partition_id"
      }
    }
  }
}

• Replace "your_api_key" with your actual Ragie API key from your dashboard.
• RAGIE_PARTITION is optional and lets you control which connected services are exposed in Cursor.

Option 2: Use a shell script 

If Option 1 throws an error, you can connect Ragie’s MCP Server via a shell script.

1. Create a file called ragie-mcp.sh:

#!/usr/bin/env bash

export RAGIE_API_KEY="your_api_key"
export RAGIE_PARTITION="optional_partition_id"

npx -y @ragieai/mcp-server

2. Make the script executable:

chmod +x ragie-mcp.sh

3. In Cursor, go to Settings → Cursor Settings → MCP Servers, and add this script as a custom MCP server. As always, replace your_api_key with your actual Ragie API key, and optionally set RAGIE_PARTITION.

You should see a green dot confirming that Ragie’s MCP Server is active. 

That's it! Cursor can now automatically pull relevant context from Google Drive, Jira, Slack, Notion, and more—right inside your IDE. 

Let’s Ask Cursor to Generate Code Based on The Connected Data

For this example, we connected Ragie’s API Docs to Cursor using Ragie’s MCP Server (in Step 1). Now, let’s ask Cursor to set up an integration for Ragie Connect in our application.

Nice, it worked!

Real-World Example: Let Cursor Audit Your Code for Compliance (SOC 2, HIPAA, etc.)

Let’s assume you’re building a new function and need to ensure it is compliant with your company’s SOC 2 or HIPAA policies. Instead of digging through Google Drive, Notion, and Slack for answers, Ragie’s MCP Server lets Cursor pull that context directly into your IDE and see if your code is compliant with your company’s policies. 

You can ask Cursor questions like:

“Does this function comply with our data retention policy?”

Cursor will surface all relevant docs and help make your code compliant powered by Ragie behind the scenes.

Here’s a quick demo showing how it works:

Use Ragie's MCP Server Beyond Cursor

MCP servers are becoming more widely used to enhance the quality of AI-powered coding assistants in IDEs like Cursor and Windsurf, helping them retrieve relevant context, understand project requirements, and streamline development workflows. But their potential goes far beyond coding.

Ragie’s MCP server can enable many more use cases, such as:

• Retrieving and summarizing project updates from Notion, Confluence, and Slack.
• Automating product feedback analysis from customer support tickets.
• Connecting AI agents to internal knowledge bases to generate instant answers.
• Powering AI workflows in non-coding environments like sales, marketing, and customer support.

We’ll be covering more use cases in upcoming guides—stay tuned!

You can try Ragie’s MCP Server for free with our developer plan. Hop on our Discord community if you have any questions; we're always happy to help. 

I hope you enjoyed learning from this guide. If you build something with Ragie’s MCP Server, tag us on Twitter or LinkedIn—we’d love to see what you ship.