Integrate a remote Infobip MCP server with an MCP Client
MCP is an open protocol that standardizes how applications provide context to AI models. The Infobip MCP server allows LLMs to use Infobip CPaaS Platform tools.
This tutorial will show you how to integrate the Infobip MCP server with an MCP Client, like Cursor or Claude desktop to use it with Infobip APIs.
Prerequisites [#prerequisites]
- Python version 3.13 upwards and
uvxinstalled. - Infobip account. If you do not have one create a free trial account.
- Infobip API key with the correct scope, e.g. for Viber use
viber-bm:message:send. Find out more about scopes in API Authorization. - An MCP Client, e.g. Cursor or Claude Desktop.
Cursor MCP Integration [#cursor-mcp-integration-steps]
Skip this section if integrating with Claude Desktop (see the section below).
- Open Cursor and navigate to Cursor Settings.
- On the left-hand side, click MCP.
- Click the +Add new global MCP server button.
- Populate the
mcp.jsonconnection to a remote MCP server with the Infobip MCP server settings (see section below). - Make sure that MCP server is enabled and you see list of tools
- Use Chat to initiate Infobip APIs.
Claude Desktop MCP Integration [#claude-desktop-mcp-integration-steps]
- Open Claude Desktop and navigate to Settings.
- Use the Developer tab to click the Edit config button. This will show you the location of the
claude_desktop_config.jsonfile in your file system. You'll use the file to add the Infobip MCP server settings. - Access the
claude_desktop_config.jsonfile and populate with the connection settings to the Infobip MCP server (see section below). - Restart Claude desktop.
- Use Chat to initiate Infobip APIs.
Infobip MCP Server settings [#infobip-mcp-server-settings]
Each Infobip API endpoint is served as a separate MCP server to narrow down the context for the AI agent in use. As such, you need to add each endpoint aka server separately, as seen below.
For the list of all supported MCP endpoints, check out the Infobip MCP documentation.