Introducing STDIO for Local Clients

STDIO (Standard Input/Output) is the primary transport for desktop AI clients. It’s fast, reliable, and requires no network configuration.

What is STDIO?

STDIO is a communication method where programs talk to each other through text streams. Your AI client launches GameCatalyst’s Conductor script, and they communicate by sending JSON messages back and forth through standard input and output pipes.

When to use STDIO:

• You’re using a desktop AI client (Claude Desktop, Cursor, Windsurf, etc.)
• You want the lowest latency and highest reliability
• You don’t need remote access from other machines
• You prefer zero network configuration

Benefits of STDIO:

• Lowest latency — Messages travel through memory, not network sockets
• Most reliable — No network issues, firewall blocks, or connection drops
• Zero configuration — No ports to open, no URLs to manage
• Automatic lifecycle — When your AI client closes, the connection closes automatically

How it works:

1. You add GameCatalyst to your AI client’s MCP configuration
2. The configuration points to the Conductor Python script
3. When your AI client starts, it launches the Conductor
4. The Conductor starts the MCP Bridge and Discovery Server (if not already running)
5. Messages flow through stdin/stdout between the AI client and GameCatalyst

Supported clients:

All desktop AI clients use STDIO: Claude Desktop, Cursor, Windsurf, GitHub Copilot, Cline, Roo Code, JetBrains AI, Claude Code, Augment Code, Gemini CLI, and Amazon Q Developer.

STDIO is the recommended transport for most users. It’s simple, fast, and works out of the box.

How STDIO works behind the scenes:

When your AI client starts, it launches GameCatalyst’s Conductor script as a child process. The Conductor communicates with your AI client through standard input (stdin) and standard output (stdout) — two text streams that every program has. Messages are formatted as JSON-RPC, a lightweight protocol for structured communication. The Conductor then connects to the MCP Bridge and Discovery Server (starting them if needed), creating the full communication pipeline.

Automatic lifecycle:

One of the nice things about STDIO is automatic lifecycle management. When your AI client starts, it launches the Conductor. When your AI client closes, the Conductor’s stdin closes, which triggers graceful shutdown. You do not need to manually start or stop anything. The connection manages itself.

Troubleshooting STDIO:

If your AI client cannot connect via STDIO, the most common causes are: Python is not on your PATH (run python --version to check), the Conductor script path is wrong in your config file (use the Configurator to generate the correct path), or JSON syntax errors in the config file (use a JSON validator). The Dashboard Activity Log shows detailed connection events that help diagnose problems.