MCP Server

The Loop MCP server lets you access feedback, projects, and developer tickets directly from your AI coding tools. Browse user feedback, generate AI-powered tickets, and update ticket status without leaving your editor.

Works with any MCP-compatible client, including Cursor, Claude Code, Windsurf, Cline, Continue, and more.

What is MCP?

The Model Context Protocol (MCP) is an open standard that allows AI assistants to access external data sources and tools. Loop provides an MCP server that exposes your organisation's feedback data to any AI assistant that supports the protocol.

How it works

AI Tool (Cursor, Claude Code, etc.)  →  MCP Server  →  Loop API V1  →  Your feedback data

1. Your AI tool spawns the Loop MCP server as a child process
2. The MCP server authenticates with your API token
3. When you ask about feedback or tickets, the AI calls MCP tools
4. The MCP server translates tool calls into Loop API requests
5. Results are returned to the AI's context

Requirements

• An MCP-compatible AI tool (Cursor, Claude Code, Windsurf, Cline, Continue, etc.)
• Node.js 18+ installed
• A Loop API token (Pro or Business plan, Owner or Developer role)

Quick start

1. Generate an API token from your Loop dashboard
2. Set up the MCP server in your tool's configuration
3. Start using the available tools

Package

The MCP server is published as an npm package:

@loop-feedback/mcp

It runs via npx so you do not need to install it globally. Your AI tool manages the process lifecycle automatically.