Arthur MCPArthur MCP
Dashboard
How it worksFeaturesDocsEarly accessGitHubDashboard

How to create User Feedback Classifier using MCP

Categorize and extract themes from user feedback. This guide creates the User Feedback Classifier prompt from Arthur's template catalog, sets up a blank MCP server, and links the prompt so any connected AI client can use it.

Steps

  1. Go to Prompt Templates and search for User Feedback Classifier.

    Open the Prompt Templates catalog

    Search the catalog for the User Feedback Classifier template

  2. Click Use template on the User Feedback Classifier card and review the prefilled name, description, and content preview.

    Open the User Feedback Classifier template with its prefilled name and content preview

    Create the User Feedback Classifier prompt in the shared library

  3. Click Create prompt — the prompt is added to your shared library.

    Start creating the MCP server that will expose the prompt

  4. Go to New server, choose Blank / Static as the source, and click Next.

    Choose the Blank / Static source — no external connection is needed for a prompt-only server

    Advance to the server details step

    Name the server after the prompt it will expose

    Create the blank MCP server

    Open the server's Prompts tab

  5. Name the server (for example, User Feedback Classifier Server) and click Create server.

    Open the prompt-library picker

  6. On the server's Prompts tab, click Add prompt, search for User Feedback Classifier, and click Add.

    Search the library for the User Feedback Classifier prompt

    Link the User Feedback Classifier prompt to the server

    Close the picker and show the linked prompt on the Prompts tab

  7. Open the server's Connect section and click Share the MCP swagger documentation — Arthur generates the public documentation link and QR code for this server.

    Open the server's Connect section, where the MCP endpoint and its documentation live

    Open the MCP swagger documentation share panel

    Show the generated swagger documentation link and QR code for this server

Confirm it worked

The User Feedback Classifier prompt appears on the server's Prompts tab, and the Connect section offers the server's MCP swagger documentation — a shareable page with setup instructions for any AI client. A connected client can now fetch the prompt through this server.

Good to know

  • Processes raw user feedback into structured categories, themes, and actionable insights.
  • Arguments detected from the template: {{count}}, {{product}}, {{feedback}} — callers fill these in when invoking the prompt.
  • Editing the prompt in the shared library updates it on every server where it's linked.

Related

Tutorial video