››

How to Connect AI Tools to Your ProProfs Knowledge Base via MCP?

MCP integration lets AI assistants such as Claude, Cursor, and Copilot read your published knowledge base content directly, so your team can get answers without leaving their AI tool. You can turn it on from Settings > Integrations > MCP in ProProfs Knowledge Base.

What Is MCP Integration?

Model Context Protocol (MCP) is an open standard that lets AI tools connect securely to external data sources like your knowledge base. In ProProfs Knowledge Base, turning on MCP exposes a secure connection endpoint that AI assistants such as Claude, Cursor, Copilot, and ChatGPT can use to browse and read your published articles. Connected tools can only access what the permissions on this page allow.

Why Use It?

MCP integration saves time whenever your team or customers already work inside an AI assistant.

Your team keeps switching tabs to search the help center. MCP lets them ask Claude or ChatGPT directly and get answers pulled from your published articles.
Support agents draft replies inside Cursor or Copilot. Connecting MCP gives their AI assistant live access to your current documentation as they type.
You need control over automated access. Read and write permissions, plus sign-in options, keep connected AI tools within the rules you set.

Example Use Cases

A support agent uses Claude Desktop to pull the latest troubleshooting steps from the knowledge base while replying to a ticket.
A developer connects Cursor IDE to reference API documentation stored in the knowledge base while writing code.
A team lead connects Microsoft Copilot Studio so an internal support bot can answer employee questions from published FAQs.

Step-by-Step Instructions

Step 1: Go to Settings > Integrations and open the MCP card.

Step 2: Toggle on Connect AI Tools to let AI assistants securely access your published content.

Step 3: Toggle on Sign in via browser, so your team can approve AI tool requests with a simple sign-in.

Note: If you leave Sign in via browser off, share an API key with the AI tool directly instead.

Step 4: Under Permissions, toggle on Read Access to let AI tools browse and read published articles.

Step 5: Toggle on Write Access only if AI tools should also create and edit articles and folders.

Note: Write Access lets connected AI tools change your published content. Turn it on only for tools your team fully trusts.

Step 6: Scroll down to Secure Connection Endpoint to view the MCP server URL for your knowledge base.

Step 7: Click Get Connection Details to open the connection popup.

Step 8: On the MCP URL / JSON tab, click Copy URL or copy the JSON Configuration block below it.

Step 9: Replace admin_api_key in the JSON with your Admin API Key from Edit Account.

Step 10: Switch to the OAuth 2.0 (assistants) tab to copy the Authorization, Token, and Revoke URLs if your AI tool needs OAuth.

Step 11: In Cursor IDE, go to Settings > Tools & MCP > Add new MCP server, paste the JSON, then save and approve access.

Step 12: In Claude (web or desktop), go to Settings > Connectors > Add custom connector, paste the MCP server URL, then sign in with your Admin API Key.

Step 13: In Microsoft Copilot / Copilot Studio, go to Tools > Add a tool > Model Context Protocol, paste the URL, and select OAuth 2.0.

Step 14: In ChatGPT (developer connectors), turn on Developer mode, add a custom connector with the MCP server URL, and sign in when prompted.

Note: If ChatGPT shows an "invalid client" or "No client ID" error, remove the connector and add it again.

Pro Tips & Interactive Settings

Tip: If your AI tool supports Dynamic Client Registration (RFC 7591), it can request a client ID automatically, so you don't need to pre-register anything.

Permissions: Read Access and Write Access can be turned on independently. Keep Write Access off until you are ready to let a connected AI tool create or edit content in your knowledge base.

Related Articles

Was this information helpful?

Thank you Your feedback helps us to continually improve our content.