mcp-analytics

MCP Analytics Suite is an intelligent analytics platform that understands natural language questions and automatically selects the right analytical approach. It guides you through data upload, method discovery, and analysis orchestration to produce interactive reports with AI-driven insights. The tools exposed by MCP Analytics include discovery of appropriate analytical methods, automated execution of analyses, data management operations, and report viewing. In practice, you describe a business question, upload or reference your dataset, and the platform handles the rest, returning a comprehensive analysis workflow and an interactive visualization suite.

Usage typically follows a flow: first you ask a question in natural language (for example, a request to analyze sales drivers or segment customers). The system uses tools.discover to select an appropriate analytical approach, then datasets.upload to securely bring your data into processing. Tools.run executes the chosen analysis configuration, and reports.view presents an interactive dashboard or report you can share. The platform supports data management (datasets.upload, datasets.list, datasets.read), core analytics tools (tools.discover, tools.run, tools.info), and reporting capabilities (reports.view, reports.search), plus platform utilities like billing and documentation access.

Prerequisites:

• Node.js and npm installed on your machine to use npx for the MCP remote runner.
• Access to the MCP Analytics API endpoint (auth0 URL provided by MCP Analytics).

1. Install Node.js and npm if you don’t have them:

   • macOS: download from https://nodejs.org
   • Windows: installer from https://nodejs.org
   • Linux: use your distribution’s package manager, e.g., sudo apt install nodejs npm

2. Verify installation:

   node -v npm -v

3. Use the MCP remote runner via npx (example provided in Quick Start):

   npx -y mcp-remote@latest https://api.mcpanalytics.ai/auth0

4. Configure your environment by adding the MCP server entry to your client’s config as shown in the examples below (depending on your integration):

• Claude Desktop

  • macOS config path: ~/Library/Application Support/Claude/claude_desktop_config.json

  • Windows config path: %APPDATA%\Claude\claude_desktop_config.json

  • Add:

    { "mcpServers": { "mcp-analytics": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"] } } }

• Cursor

  • Place in .cursor/config.json in your project root:

    { "mcpServers": { "mcp-analytics": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"] } } }

• VS Code (Continue Extension) - config at ~/.continue/config.json:

  { "models": [{ "provider": "anthropic", "model": "claude-3-5-sonnet", "mcpServers": { "mcp-analytics": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"] } } }] }

• Claude Code

  • Add to claude_code_config.json:

    { "mcpServers": { "mcp-analytics": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://api.mcpanalytics.ai/auth0"] } } }

Notes and tips:

• The MCP remote runner uses OAuth2 with the provided auth URL; ensure your credentials are configured in the MCP Analytics account to access the API.
• Data uploaded via datasets.upload is handled securely; consider using encryption in transit (TLS) and ephemeral processing as described in the security docs.
• If the remote runner cannot reach https://api.mcpanalytics.ai/auth0, verify network egress rules and the correctness of the Auth0 URL.
• For large datasets, prefer incremental uploads or chunked transfers if supported by your client integration.
• The tooling supports a range of methods (discover, run, info) and reporting (view, search). You can combine these to build interactive dashboards and shareable insights.