Managing n8n workflows through ChatGPT with a custom MCP server

At X Mind Solutions, we developed and deployed a custom MCP server designed to work with OpenAI’s MCP (Model Context Protocol) framework. This integration lets us manage the n8n system in our test environment through ChatGPT without opening n8n’s own interface. Using natural-language requests, we can create new workflows, modify existing ones, test them, and trigger them as needed. In this way, we are making the chat interface part of automation management.

In the architecture we developed, the MCP server acts as the connection layer enabling communication between ChatGPT and n8n. Users express their needs in the conversation, while workflow operations are carried out in n8n. The distinction here is that the interaction is not limited to starting a preconfigured automation. Creating a new workflow on request, modifying an existing one, and then testing it are all possible within the same interaction.

One application of this approach is recruitment. Our example scenario addresses various automation needs involving personnel files, CVs, meeting notes, and intern folders in Drive. Tasks such as creating an initial record, adding meeting notes, and tracking the internship process can be turned into separate n8n workflows. Users can define these needs by talking to ChatGPT and initiate tests of the resulting workflows in the same conversation.

From a user experience perspective, the key change is fewer switches between interfaces when moving from describing a business need to managing its automation. Rather than opening n8n and performing operations there, users can submit requests to create and test workflows through the conversation. This brings natural language into workflow management. However, a workflow running in a test environment is not the same as being ready for enterprise use; process requirements must be assessed separately.

With this development, X Mind Solutions brings together AI-powered conversation, system integration, and automation workflows. Our current implementation covers the n8n system in our test environment; we do not present the capabilities achieved here as a guarantee of results across all enterprise environments. The key development is that ChatGPT can be used to manage not only workflow execution, but also workflow creation, modification, and testing.