> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://docs.middesk.com/build/mcp-server/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.middesk.com/_mcp/server. # Connect to Middesk's MCP server > Connect AI models and agents to Middesk data using the Model Context Protocol (MCP) server. The Model Context Protocol (MCP) server enables compatible AI models and agents to securely access Middesk data. The server supports both Server-Sent Events (SSE) and Streamable HTTP transports with OAuth 2.1 authentication. Middesk provides two transport endpoints: * **SSE**, `https://mcp.middesk.com/sse` * **HTTP**, `https://mcp.middesk.com/mcp` ## Install the MCP server for Claude Installation steps vary on the version of Claude. #### Claude Team or Enterprise (Claude.ai) 1. Navigate to **Settings** > **Integrations**. 2. Click **Add more**. 3. Enter **Middesk** as the integration name. 4. Enter the URL: `https://mcp.middesk.com/sse` 5. Enable the tools in new chats. #### Claude Desktop (Free or Pro) 1. Open the configuration file at `~/Library/Application Support/Claude/claude_desktop_config.json`. 2. Add the following configuration: ```json { "mcpServers": { "middesk": { "command": "npx", "args": ["-y", "mcp-remote", "https://mcp.middesk.com/sse"] } } } ``` 3. Restart Claude Desktop. ## Install the MCP server for other MCP-compatible tools The `mcp-remote` module provides backwards compatibility with clients that do not natively support remote MCP. This enables a wider range of AI tools and agents to connect to Middesk data. ```bash npx -y mcp-remote https://mcp.middesk.com/sse ``` ## Configure with an API key For clients that support Streamable HTTP, you can configure the MCP server directly with an API key instead of using OAuth: ```json { "mcpServers": { "middesk": { "url": "https://mcp.middesk.com/mcp", "headers": { "Authorization": "Bearer mk_live_your_api_key" } } } } ``` Replace `mk_live_your_api_key` with your Middesk API key. See [Get and use your API keys](/build/api-keys) to obtain your key. > **Get a demo** > > Contact your account manager or [contact sales](https://www.middesk.com/contact-sales) to inquire about access. > Connect AI models and agents to Middesk data using the Model Context Protocol (MCP) server.