Skip to main content

How to Configure MCP Integrations

Configure Model Context Protocol integrations in ArcPrime to connect external AI tools and workflows

Written by Jonathan Liu

This article explains how to connect external AI assistants to your ArcPrime patent data using the Model Context Protocol (MCP) server.

📷 The MCP Server settings page showing the organization selector, the server URL with copy button, and setup instruction sections for Claude Desktop, Claude Code, GitHub Copilot, ChatGPT, Codex CLI, and Gemini CLI

Prerequisites

  • You have admin or settings access in your ArcPrime workspace.

  • You have an account with at least one supported AI assistant (Claude, Copilot, ChatGPT, or Gemini).

Steps

Find Your MCP Server URL

  1. Open Settings from the sidebar.

  2. Select MCP Server from the settings menu.

  3. If your account has access to multiple organizations, select the correct one from the Organization dropdown in the top-right corner.

  4. Your unique server URL appears in the Server URL card. Click Copy to copy it to your clipboard.

Connect Claude Desktop

  1. In Claude Desktop, open Settings and select Connectors.

  2. Click Add custom connector.

  3. Paste your ArcPrime MCP server URL.

  4. Click Add. A browser window opens for you to log in with your ArcPrime credentials.

Connect Claude Code

  1. Run the following command in your terminal, replacing the URL with your server URL from ArcPrime.

  2. On first use, you are prompted to authenticate through your browser.

Connect GitHub Copilot

  1. In VS Code, open the Command Palette.

  2. Search for MCP: Add Server.

  3. Select HTTP as the server type.

  4. Paste your ArcPrime MCP server URL and enter arcprime as the server name.

Connect ChatGPT

  1. In ChatGPT, open Settings and select Apps & Connectors.

  2. Enable Developer Mode under Advanced settings if not already active.

  3. Click Create, enter a name (e.g., "ArcPrime"), and paste your server URL.

Note: The MCP server is rate-limited to 10 requests per minute per workspace. If you exceed this limit, requests return a retry message.

Related Articles

Did this answer your question?