Overview

We are proud to introduce our Model Context Protocol (MCP) Server. This new functionality allows developers to bridge the gap between our API documentation and AI-assisted development tools like Claude, ChatGPT, Cursor, and GitHub Copilot.

Instead of manually searching through documentation pages, you can now connect your AI assistant directly to our live documentation to get instant, contextual answers and accurate code snippets.

Why use the MCP?

The traditional "read-and-research" method of integration is evolving. By using our MCP server, you benefit from:

• Zero-Search Integration: Ask questions in natural language (e.g., "How do I sync tax rates?") and get immediate answers.
• Contextual Code Generation: Receive code examples that are tailored to your specific query and reflect the latest API schema.
• Eliminated Learning Curves: New team members can start building immediately without mastering our entire documentation structure first.
• Real-Time Accuracy: The AI references the live documentation, ensuring you never build against deprecated endpoints.
• Reduced Context Swithching: Stay in you IDE or AI assistant without jumping between docs, browser tabs, and code - keeping you in flow and shipping faster. 

How it works

The MCP is an open standard that enables AI models to access external data sources securely. By hosting an MCP server, we provide a “translator” between our technical documentation and your LLM-bases tools. 

For a partner developer looking to accelerate their workflow, the experience looks like this:

• Configuration: In your AI tool of choice (e.g., Claude Desktop, Cursor, ChatGPT), add the MCP server URL to your configuration.
• Indexing: The AI tool connects to the server and indexes the available documentation automatically.
• Querying: You ask questions in natural language — for example, “How do I provision company data; dimensions and users?”
• Guided Response: The AI responds with answers, code samples, and guidance grounded in the current documentation, including references back to the relevant doc pages.
• Conversational Iteration: You refine the integration, ask follow-up questions, and generate code—all without leaving your development environment.

Example Scenarios

• New Partner Onboarding: *Prompt: “How do I integrate with the Admin API to provision users?”
  Result: The AI walks through the relevant flow end-to-end.
• Code Generation: *Prompt: “Give me a TypeScript example for calling the Dimensions endpoint”
  Result: The AI generates a working sample using the latest schema.
• Troubleshooting: *Prompt: “Why might I be getting a 404 from this endpoint?”
  Result: The AI explains potential causes based on documented authentication and authorization rules.
• Discovery: *Prompt: “What APIs are avaliable for managing organizations?”
  Result: The AI lists relevant endpoints with short descriptions and links. 

Getting Started

To begin using the AI Assistant, follow there steps: 

1. Locate the MCP URL: The config is located in the Developer Portal
2. Configure Your Tool: 
   In Cursor: Go to Settings > Models > MCP and add the URL
   In Claude Desktop: Add the configuration to your claude_desktop_config.json file.
3. Start Prompting: Once connected, your AI is ready to assist with any specific development tasks. 

How it looks

Additional information

• The MCP server only exposes public API documentation - No partner, customer, or tenant data is accessible through it. 
• Coverage: The MCP server reflects whatever is currently published in the developer documentation. As docs are updated, MAP responses update automatically. 
• AI-native developer experience. Faster integrations, lower onboarding friction and a modern use of powerful AI tools.