How to Connect Any MCP Client to a Remote MCP Server

This guide provides a universal checklist for connecting any MCP client to a remote MCP server. While the exact configuration steps vary by client, the core principles are the same. We'll use Corcava as the worked example throughout.

Universal Connection Checklist

Regardless of which MCP client you're using, you'll need to complete these steps:

Step 1: Get Your API Key

Before connecting, you need an API key from the MCP server. For Corcava:

1. Log in to your Corcava workspace
2. Navigate to Settings → Integrations
3. Find the Public API section
4. Click Add API Key
5. Copy the generated key immediately (it's only shown once)
6. Store it securely - you'll need it for the next steps

Step 2: Locate Your Client's Config File

Each MCP client stores its configuration in a different location. Here are the common locations:

For detailed, client-specific instructions, see our setup guides:

• Claude Desktop Setup Guide
• Cursor Setup Guide
• Windsurf Setup Guide
• Continue Setup Guide

Step 3: Add the Server Endpoint URL

The endpoint URL is the address where your MCP client will connect to the server. For Corcava, this is:

This URL should be:

• HTTPS: Always use secure connections (never HTTP)
• Exact: Copy the URL exactly - no trailing slashes
• Accessible: Ensure your network can reach this endpoint (check firewalls/proxies if needed)

Step 4: Add the Authorization Header

Most MCP clients require you to specify authentication headers. For Corcava, you need to add an Authorization header with your API key.

Example Configuration Snippets

{
  "mcpServers": {
    "corcava": {
      "url": "https://app.corcava.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

{
  "mcpServers": [
    {
      "name": "corcava",
      "transport": {
        "type": "sse",
        "url": "https://app.corcava.com/mcp",
        "headers": {
          "Authorization": "Bearer YOUR_API_KEY"
        }
      }
    }
  ]
}

Step 5: Save and Restart

After adding the configuration:

1. Save the configuration file (if editing a file)
2. Validate JSON syntax (use a JSON validator if unsure)
3. Completely restart your AI client - closing and reopening is required
4. Wait for the client to initialize - this may take a few seconds

Step 6: Verify Tools Appear

After restarting, verify that your AI client can see the MCP server's tools. Here's how to check:

Expected Tool List

For Corcava, you should see these tools (and more):

• list_tasks - Search and filter tasks
• get_task - Get task details
• create_task - Create new tasks
• update_task - Update task details
• list_projects - List projects
• start_time_tracking - Start timer
• And more... (see complete tools reference)

Step 7: First Test Prompt

Once tools are visible, test with a simple read-only operation:

Example: Testing with Corcava

Troubleshooting Connection Issues

If tools don't appear or you get errors, check these common issues:

Client-Specific Guides

For detailed, step-by-step instructions for each client, see our dedicated setup guides: