Skip to content

Claude Code

Claude Code

Use the skill path first before dropping to raw MCP setup.

  • Claude Code installed locally
  • Access to the Google or GitHub sign-in you want to connect when Claude Code opens browser approval
Section titled “Recommended: install the Claude Code skill”
Terminal window
npx skills add agent-analytics/skills

This path teaches Claude Code how to set up tracking, query analytics, and run experiments. It uses browser approval, agent sessions, and human-owned upgrade handoffs by default.

Ask Claude Code:

  • Set up Agent Analytics for this project. Install it here if needed. Open the browser for me or give me a login link, then wait. I will sign in with Google or GitHub, approve it, and paste back any finish code if you need it. Then create the project, add tracking and key events, and verify the first event.
  • List my Agent Analytics projects
  • How is my-site doing this week?
  • What are the top pages for my-site this week?

If the skill is working, Claude Code should be able to reach your account without asking you to hand-roll HTTP requests.

If you have not created your first real project yet, continue with First Project in 5 Minutes.

Lower-level fallback: add only the MCP server

Section titled “Lower-level fallback: add only the MCP server”

If you want the MCP server without the skill layer:

Terminal window
claude mcp add agent-analytics --transport http https://mcp.agentanalytics.sh/mcp

Use --transport http. The hosted MCP server is not configured for legacy SSE transport.

  • Make sure Claude Code completes browser approval with the same GitHub or Google account you want connected to Agent Analytics.
  • If the skill pauses on approval, complete the browser sign-in and let Claude Code continue.
  • If you intentionally use a compatibility API key for direct HTTP debugging, confirm the environment exposes AGENT_ANALYTICS_API_KEY.
  • If the MCP command fails, verify you used --transport http.