The Wonderflow MCP Connector gives your AI assistant direct access to your scope’s catalog using live Wonderflow data. Once connected, the assistant operates with your user identity, permissions, and scope visibility — no separate API credentials required. 🚀

​
🤖 What is MCP?

The Model Context Protocol is an open standard that allows AI applications to connect to external data sources and tools in a secure, standardized way. Think of it as a universal adapter between AI assistants and the services they need to access.

With the Wonderflow MCP Connector, your AI assistant can:

  • 🔍 Discover topics — find out what topics your catalog has identified recently
  • 📊 Query analytics — ask natural-language questions about your customer feedback data
  • 💬 Browse feedback — retrieve, filter, and analyze feedback entries

​
🌐 Endpoint

MCP server URLClient ID
https://catalog-mcp.wonderflow.ai/mcplyLq7SwG6IcGDc9zY52xqTyRcLkUKIo8

Most clients only need the server URL. Some clients, such as Copilot Studio, also need the client ID and secret in their configuration: the setup guide of each client tells you when.

​
🔑 Who can connect?

Any Wonderflow user with a Wonderboard account, once the MCP Connector has been enabled for your account. Authentication uses your existing Auth0 login, with no extra credentials and no service accounts.

The MCP Connector is a paid feature and must be enabled for your account. Once it is enabled, Wonderflow shares any additional configuration parameters your client may need based on the client application you are using, for example the client secret required by Microsoft Copilot Studio. To get access, contact your Wonderflow representative.

​
⚙️ How it works

  1. Your AI assistant connects to the Wonderflow MCP server endpoint
  2. You authenticate via Auth0 using your existing Wonderflow credentials
  3. The assistant can now call Wonderflow catalog tools scoped to your identity and permissions

​
🗂️ Catalogs and scopes

The MCP Connector works on one catalog at a time: the scope you are currently using in the Wonderboard. Whatever scope is selected there is the catalog your assistant sees.

To move to a different catalog, select another scope in the Wonderboard with the scope selector, then go back to your MCP client and open a new chat. The new chat uses the scope you selected.

​
Switching catalog from the chat

If you have a Wonderflow admin account, you can also switch catalog directly from the chat, without opening the Wonderboard. This is handy, for example, during demos. Ask your assistant to switch and give it the scope ID of the catalog. You can see the scope ID by hovering over the catalog in the Wonderboard.

“Switch to the catalog with scope ID <scope-id>”

Behind the scenes the assistant uses the list_catalogs and switch_catalog tools.

Always provide the scope ID when you switch from the chat, so the assistant selects exactly the catalog you mean.

​
📚 Getting started

Authentication

Learn how OAuth 2.1 authentication works with the MCP Connector.

MCP Clients

Set up your preferred AI assistant to work with Wonderflow.

Tools

Explore all the available tools — analytics, discovery, products, and more.

​
🧩 Supported clients

Any client that supports OAuth 2.1 + Streamable HTTP transport should work. Below are the ones that have been tested.

Most clients discover and register with our authorization server automatically. Clients that work differently, such as Microsoft Copilot Studio, need a manual OAuth configuration with parameters provided by Wonderflow once the connector is enabled for your account.

Claude Desktop

Best for conversational exploration of your catalog — ask questions in natural language from Anthropic’s desktop app.

ChatGPT

Best for conversational exploration of your catalog from OpenAI’s ChatGPT, added as a custom MCP plugin.

Microsoft Copilot Studio

Best for teams building agents in the Microsoft ecosystem. Requires a manual OAuth setup with parameters provided by Wonderflow.

Cursor

Best for developers who want Wonderflow data alongside their code, right inside their IDE.

Gemini CLI

Best for terminal-first workflows using Google’s Gemini models.

MCP Inspector

Best for developers who need to debug tool calls, inspect schemas, and test raw MCP responses.

Authentication