Metaplay works exceptionally well with AI coding assistants. The SDK ships as source code with comprehensive documentation and full sample projects, giving AI tools the context they need to reason about your codebase and architecture.

Why Metaplay + AI Works Well

While the latest frontier models are incredibly capable on their own, they really become useful when given access to your specific codebase and environment.

• Source code access - The SDK is available as C# and TypeScript source, not compiled binaries.
• Server logs and metrics - Secure access to real-time observability into your game's performance and issues.
• Comprehensive docs - Architecture, patterns, and operational guidance are documented in depth.
• Fast feedback loops - Metaplay includes tests and tooling that help keep agent output grounded.

Which AI Assistant Should You Use?

You can think of AI assistants as belonging to three different categories:

1. Casual interest

   Use our Metaplay Docs Custom GPT to chat about Metaplay concepts and patterns. Works on your phone and web and is free to use.

2. Personal assistant

   To get more specific help, you can install our Metaplay Docs MCP and Metaplay Portal MCP to give any agentic AI assistant access to our documentation and your project's live status. The agent can then also manage your Metaplay portal for you.

   Our preferred agents are Claude and ChatGPT. After the initial setup, you can use their mobile/web/desktop apps to connect to your own game and have them help you with planning. This is especially powerful when you connect the agents to more services, like your task tracking.

3. Coding assistant

   For actual coding tasks, you can you can install our Metaplay Docs MCP and Metaplay Portal MCP to give the coding agent access to example implementations and let it dive into production logs to understand issues.

   Our team generally prefers Claude Code for programming, but other agentic coding harnesses like Codex and Windsurf work well too.

Choose by Task

• "I need a quick explanation or where-to-start answer." - Use Metaplay Docs Custom GPT in ChatGPT.
• "I want my coding agent to read docs + SDK source while planning." - Use Metaplay Docs MCP in Claude Code or Codex.
• "I need to inspect production issues, manage roles, or query environment statuses." - Use Metaplay Portal MCP in Claude or ChatGPT.

Note:

AI assistants are extremely useful, but they still make assumptions. Always verify destructive changes, production-impacting actions, and security-sensitive decisions.

MCP... What?

MCP stands for Model Context Protocol. It's a protocol that allows AI assistants to access external tools and data sources, like our developer portal and documentation.

To make an AI assistant be aware of Metaplay's tools, we have created "MCP servers" that the AI assistant can connect to. These servers then provide the AI assistant a list of tools and other resources it can call on. You can control the AI's access to these tools one-by-one and we can keep adding new ones over time.

Since MCP is not a great term, many companies have started calling them "connections" or "integrations" in their apps. Under the hood, they're all MCP servers.

Coming Soon

We are working on a dedicated Metaplay programming assistant distributed as a Claude Code plugin.

This upcoming assistant is designed to combine:

• Documentation and SDK context from Metaplay Docs MCP.
• Live operational context from Metaplay Portal MCP.
• Ability to work on actual coding tasks in your project.