Transports

The MCP server supports two transport modes, selected by the MCP_TRANSPORT environment variable.

HTTP (production)

The default for production. Runs an Express 5 server with Streamable HTTP transport.

• URL: https://mcp.campaignstack.io/mcp
• Auth: Required — Authorization: Bearer cs_...
• Protocol: JSON-RPC 2.0 over HTTP, with SSE for streaming

Session lifecycle

1. Initialize — POST /mcp with initialize method. Server returns a session ID in the mcp-session-id header.
2. Tool calls — POST /mcp with tools/call method, include session ID header.
3. Streaming — GET /mcp with session ID for SSE event stream.
4. Close — DELETE /mcp with session ID.

Health check

curl https://mcp.campaignstack.io/health

Returns server version, status, and the complete list of available tools (no authentication required).

stdio (local development)

For local development and testing. No authentication — all scopes are granted.

MCP_TRANSPORT=stdio npx tsx src/index.ts

The server reads JSON-RPC messages from stdin and writes responses to stdout. This is how MCP clients like Claude Desktop and Cursor connect to local servers.

Claude Desktop config (stdio)

{
  "mcpServers": {
    "campaignstack": {
      "command": "npx",
      "args": ["tsx", "src/index.ts"],
      "cwd": "/path/to/CampaignStack/app/mcp-server",
      "env": {
        "MCP_TRANSPORT": "stdio",
        "CONVEX_URL": "https://your-deployment.convex.cloud"
      }
    }
  }
}

Testing with MCP Inspector

The MCP Inspector is a visual tool for testing MCP servers:

npx @modelcontextprotocol/inspector npx tsx src/index.ts

This launches a web UI where you can browse tools, call them interactively, and inspect responses.