Skip to main content
ClosedLoop AI is built for the agentic stack. Connect Claude Code, Cursor, or any MCP-compatible AI tool to your workspace, and your AI dev agent can query product intelligence, explore revenue-scored opportunities, and trace insights back to real customer quotes — all without leaving your IDE.

​
Setup

​
Claude Code

Step 1: Add the MCP server: 
# US workspace
claude mcp add --transport http closedloop-ai https://mcp.closedloop.sh

# EU workspace
claude mcp add --transport http closedloop-ai https://eu.mcp.closedloop.sh
Step 2: Restart Claude Code, then type /mcp to see available servers. Select closedloop-ai. Step 3: Claude Code will show an authorization link. Click it — your browser opens the ClosedLoop AI sign-in page. Log in (or sign up), select your workspace, and click Connect. Step 4: You’re connected. Start asking questions about your product intelligence.

​
Cursor

Add to your Cursor MCP settings (.cursor/mcp.json): 
{
  "mcpServers": {
    "closedloop-ai": {
      "url": "https://mcp.closedloop.sh"
    }
  }
}
For EU workspaces, use https://eu.mcp.closedloop.sh instead. Cursor will prompt you to authorize on first use.

​
Other MCP Clients

Any client supporting MCP HTTP transport with OAuth 2.1 can connect:
RegionServer URL
UShttps://mcp.closedloop.sh
EUhttps://eu.mcp.closedloop.sh
Authentication is OAuth 2.1 with PKCE, auto-discovered via /.well-known/oauth-authorization-server.

​
Example Queries

Once connected, just ask your AI tool naturally:
  • “What are the top customer pain points?”
  • “Show me deal blockers for enterprise customers”
  • “What are US entertainment companies frustrated about?”
  • “How has insight volume changed over the last 3 months?”
  • “Which opportunities have the most revenue at risk?”
  • “What are competitors doing better than us?”
  • “Build a PRD for calendar management based on customer evidence”

Tools Reference

Complete reference for all 12 available MCP tools