Other Clients

Set up ContextPin with other MCP-compatible AI tools.

ContextPin works with any AI tool that supports the Model Context Protocol. This guide provides general instructions for connecting to other MCP clients.

Generic Setup Instructions

The setup process is similar for all MCP clients:

  1. Find your client's MCP configuration location
  2. Add the ContextPin server configuration
  3. Restart the client
  4. Verify the connection

ContextPin MCP Configuration

Use this configuration template and adjust paths for your platform:

macOS

{
  "contextpin": {
    "command": "/Applications/ContextPin.app/Contents/MacOS/contextpin-mcp",
    "args": ["--workspace", "YOUR_WORKSPACE_ID"]
  }
}

Windows

{
  "contextpin": {
    "command": "C:\\Program Files\\ContextPin\\contextpin-mcp.exe",
    "args": ["--workspace", "YOUR_WORKSPACE_ID"]
  }
}

Linux

{
  "contextpin": {
    "command": "/opt/contextpin/contextpin-mcp",
    "args": ["--workspace", "YOUR_WORKSPACE_ID"]
  }
}
Get your workspace ID from ContextPin: Settings → MCP → Copy Configuration

Known MCP Clients

These clients are known to support MCP:

Client Config Location Notes
Claude Desktop ~/Library/Application Support/Claude/claude_desktop_config.json Full Guide
Claude Code .mcp.json in project or home directory Full Guide
Cursor Settings JSON or UI Full Guide
Windsurf Settings JSON Full Guide
Zed ~/.config/zed/settings.json MCP support in development
Continue ~/.continue/config.json VS Code extension

Configuration Formats

Different clients may expect slightly different configuration formats:

Nested under "mcpServers"

{
  "mcpServers": {
    "contextpin": {
      "command": "...",
      "args": ["..."]
    }
  }
}

Under "mcp.servers"

{
  "mcp.servers": {
    "contextpin": {
      "command": "...",
      "args": ["..."]
    }
  }
}

As Array

{
  "mcp": {
    "servers": [
      {
        "name": "contextpin",
        "command": "...",
        "args": ["..."]
      }
    ]
  }
}

Check your client's documentation for the expected format.

Transport: stdio

ContextPin's MCP server uses the stdio transport, which is the most common and widely supported:

  • Communication happens over stdin/stdout
  • No network ports required
  • Works locally on your machine

If your client asks for transport type, select "stdio" or "command".

Testing the Connection

To verify ContextPin is properly connected:

  1. Start your AI client
  2. Ask it to list available MCP tools or servers
  3. You should see "contextpin" listed
  4. Try: "List my ContextPin contexts"

Manual Testing

You can test the MCP server directly from the command line: 

# macOS
/Applications/ContextPin.app/Contents/MacOS/contextpin-mcp --workspace YOUR_WORKSPACE_ID

# The server will wait for MCP JSON-RPC messages on stdin

If it starts without errors, the server is working correctly.

Common Issues

Command Not Found

  • Verify ContextPin is installed at the expected location
  • Use absolute paths in configuration
  • Check file permissions

Invalid JSON

  • Validate your configuration with a JSON linter
  • Check for trailing commas
  • Ensure proper quoting

Workspace ID Error

  • Get the correct ID from ContextPin Settings → MCP
  • Don't use placeholder text
  • Check for copy/paste errors

No Response

  • Make sure ContextPin is running
  • Check client's output/logs for errors
  • Try restarting both ContextPin and the client

Building Custom Integrations

If your tool doesn't support MCP directly, you might be able to:

  • Use an MCP client library to build a plugin
  • Export contexts as files and use file-based context
  • Request MCP support from the tool's developers

Resources

Next Steps