What is MCP?
The Model Context Protocol (MCP) is an open standard that enables AI applications to securely connect to external data sources and tools. Logic implements MCP as a server, allowing you to connect AI platforms like Claude, Codex, and ChatGPT directly to your Logic documents. With Logic’s MCP server, you write your AI rules and workflows once in Logic, and any MCP-compatible tool can use them.How Logic Implements MCP
Logic provides MCP servers at the organization level. Each MCP server:- Exposes your Logic documents as tools that AI assistants can call
- Manages document access through flexible scope controls
- Supports async execution for long-running workflows
- Handles authentication automatically
Key Features
Secure Access Control
Choose which documents to expose through each MCP server:- Full Organization: All documents in your organization are available
- Limited Scope: Only selected documents are accessible
Async Task Support
Logic MCP servers support asynchronous task execution, meaning AI assistants can start long-running Logic workflows and check back for results without blocking. This enables:- Parallel execution of multiple Logic documents
- Handling of complex, time-intensive workflows
- Better resource utilization
Platform Compatibility
Logic’s MCP implementation works with major AI platforms:- Claude (desktop and CLI)
- Codex CLI
- ChatGPT
- Any other MCP-compatible client
Use Cases
Code Style Validation: Create a Logic agent that validates code against your company’s style guide. Connect it via MCP to Claude Code, and automatically check your code as you write. Data Analysis Pipeline: Build a Logic document that analyzes data and generates reports. Connect it to ChatGPT so you can run analyses conversationally. Multi-Step Workflows: Design complex Logic workflows that AI assistants can trigger and monitor, enabling sophisticated automation without custom integration code.Getting Started
Ready to connect your Logic documents to AI platforms?Start with the Quick Start guide to create your first MCP
server and connect it to an AI platform.

