Puzzle MCP (Model Context Protocol) connects your operational blueprint in Puzzle to AI tools like Claude or ChatGPT. Once connected, your AI agents can read your entire operational architecture — processes, teams, tools, and data models — and understand how they all connect. Instead of working from generic knowledge, your AI works from how your specific business actually runs.

Your agents can answer detailed questions about how any process works, understand upstream and downstream dependencies, see which tools and roles are involved at every step, and know where deterministic automations already handle things. They can also propose changes back to your blueprint — documenting what changed, why, and which steps are affected — so your operational architecture stays current.

The more complete your blueprint, the more useful your AI agents become. We recommend mapping your core processes, team structure, tools, and data models first. You can build your blueprint with Nova AI in minutes — then when you connect via MCP, your agents have full context from day one.

Puzzle MCP is a standard Model Context Protocol server, so it works with any MCP-compatible client. That includes Claude, ChatGPT, Cursor, and any other AI tool or multi-agent framework that supports MCP. The same connector serves as the shared operational context layer across your entire AI stack.