Optimizely Analytics MCP server quickstart

  • Updated

Connect your agent to Optimizely Analytics with the following steps:

Prerequisites

  • An Optimizely account with Opal enabled and connected to at least one Optimizely Analytics app.
  • One of the following AI clients installed:
    • Claude Code (CLI)
    • Claude Desktop
    • Claude.ai
    • Cursor
    • Visual Studio Code
    • Windsurf

Add the MCP server to your AI client

Add the Optimizely Analytics MCP server URL to your client's MCP configuration:

https://analytics.mcp.opal.optimizely.com/mcp

Where you add this depends on your client:

  • Claude Code – Run claude mcp add --transport http optimizely-analytics https://analytics.mcp.opal.optimizely.com/mcp.
  • Claude Desktop – Go to Settings > Developer > Edit Config and add the server to claude_desktop_config.json.
  • Claude.ai – Go to Settings > Connectors > Add custom connector.
  • Cursor – Go to Settings > Tools & MCPs > Add a Custom MCP Server, or edit .cursor/mcp.json in your project root.
  • Visual Studio Code – Open MCP settings in your GitHub Copilot configuration, or edit settings.json.
  • Windsurf – Go to Settings > MCP Servers, or edit .windsurf/mcp.json in your project root.

For full per-client steps, see Install Optimizely Analytics MCP server.

Authenticate with Optimizely

The first time your AI client calls the Optimizely Analytics MCP server, an Open Authorization (OAuth) flow opens in your browser.

  1. Log in with your Opti ID credentials (Optimizely's unified identity system, the same login you use for the Optimizely UI).
  2. Connect to your Opal instance.
  3. Authorize the MCP connection.

When the OAuth flow finishes, your AI client confirms the connection.

You do not need to create a separate API token or manage credentials manually. The MCP server uses your Optimizely identity and respects your existing account permissions.

Your authentication session is time-limited. When it expires, your AI client lets you know that you need to sign in again. The exact prompt varies by client, but you can complete the OAuth flow again by following the prompt.

Run your first query

After you authenticate, your agent has access to your Optimizely Analytics data. To run your first query, enter: List my Optimizely Analytics apps.

Your agent calls oa_list_apps, retrieves the apps you can access, and returns names and IDs. If you only have one app, or if you have set a default app, your agent resolves it automatically and you can move straight to your next prompt.

Verify the connection works

A working connection lets you:

  • See Optimizely Analytics tools in your AI client's available tools list (look for tools prefixed with oa_).
  • Get a response to List my Optimizely Analytics apps that includes real apps from your account.
  • Ask follow-up questions about the returned data without re-authenticating.

If the server does not respond or the oa_ tools are not listed, see the Troubleshoot section of the Install Optimizely Analytics MCP server article.

What to try next

When your connection works, try these prompts to explore the full range of capabilities. Replace placeholder text such as EXPERIMENT_NAME with the name of one of your real experiments or dashboards.

  • Build an exploration – Show me a funnel from sign-up to first export over the last 30 days.
  • Analyze an experiment – Analyze the EXPERIMENT_NAME experiment.
  • Build a dashboard – Create a growth dashboard with daily active users (DAU), weekly retention, and the signup funnel.
  • Run a saved metric – Run the DAU metric for the past week.

Next steps