MCP Overview

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI clients call external tools and read resources over a single HTTP+SSE or stdio transport. It is the same mechanism Claude uses to call web search, run code, and read files.

What Autousers exposes

• 39 tools across four categories: templates (6), evaluations (14), autousers (15), settings (4)
• 3 resource URIs: autousers://evaluation/{id}, autousers://template/{id}, autousers://autouser/{id}
• 5 canned prompts: evaluate-url, compare-designs, analyze-results, calibrate-autouser, triage-low-agreement

Why use it?

• Run UX evaluations directly from your IDE without switching tabs
• Ask Claude to summarise evaluation results or compare two designs
• Automate evaluations in CI pipelines using API key auth
• Calibrate and manage autousers via natural-language prompts

Server URL

https://mcp.autousers.ai/mcp

Primary MCP endpoint (also reachable at https://app.autousers.ai/mcp)

Connect your client

• Claude.ai / Claude Desktop — see the Claude.ai article
• Claude Code CLI — see the Claude Code article
• Cursor — see the Cursor article
• VS Code + GitHub Copilot — see the VS Code article
• ChatGPT — see the ChatGPT article
• Codex CLI — see the Codex CLI article
• Gemini CLI — see the Gemini CLI article
• Cline, Zed, Continue, Goose — see the stdio bridge article