Skip to main content

Connecting AI Clients

To use WIXY's AI features, you need to connect an MCP-compatible client to the WIXY management port.

1. Claude Desktop​

To add WIXY to Claude Desktop, modify your configuration file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json

Add WIXY to the mcpServers object:

claude_desktop_config.json
{
  "mcpServers": {
    "wixy": {
      "command": "curl",
      "args": ["-s", "http://localhost:8080/wixy/mcp"]
    }
  }
}
tip

Since WIXY uses SSE over HTTP, we use curl to pipe the event stream to Claude.

2. Cursor / IDEs​

For IDEs like Cursor or Windsurf:

  1. Open Settings > Features > MCP.
  2. Click + Add New MCP Server.
  3. Name: Wixy.
  4. Type: sse.
  5. URL: http://localhost:8080/wixy/mcp.

3. MCP Inspector (Debugging)​

The MCP Inspector is the best way to verify that your tools are correctly exposed.

npx @modelcontextprotocol/inspector http://localhost:8080/wixy/mcp

This will launch a web interface where you can:

  • List all available tools (listStubs, createStub, etc.)
  • Run tools manually to see their outputs.
  • Inspect the JSON-RPC traffic.

4. Custom Python/Node Clients​

You can connect your own custom agents using the official SDKs:

Python Example
from mcp import ClientSession
from mcp.client.sse import sse_client

async with sse_client("http://localhost:8080/wixy/mcp") as (read, write):
    async with ClientSession(read, write) as session:
        await session.initialize()
        tools = await session.list_tools()
        print(f"Discovered {len(tools.tools)} tools")