Analytics MCP server
{
"mcpServers": {
"mcpanalytics": {
"type": "http",
"url": "https://api.mcpanalytics.ai/auth0"
}
}
}{
"mcpServers": {
"mcpanalytics": {
"type": "http",
"url": "https://api.mcpanalytics.ai/auth0"
}
}
}{
"mcpServers": {
"mcpanalytics": {
"type": "http",
"url": "https://api.mcpanalytics.ai/auth0"
}
}
}{
"servers": {
"mcpanalytics": {
"type": "http",
"url": "https://api.mcpanalytics.ai/auth0"
}
}
}{
"mcpServers": {
"mcpanalytics": {
"type": "http",
"url": "https://api.mcpanalytics.ai/auth0"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.7
What it does
The MCP Analytics Suite exposes tools to build, run, and query statistical analyses on CSV data or live data sources. It connects to Claude, Cursor, and other MCP clients, letting you commission custom analyses (t-tests, regression, forecasting, segmentation, and more), store them in a private library, and rerun them on fresh data. Built for analysts, researchers, and anyone who needs citable, reproducible statistical reports.
- Upload a CSV and ask a question to get an instant statistical report with charts, key figures, and methodology in 2–45 minutes depending on depth tier chosen.
- Rerun a previously commissioned analysis on new data by querying your private library from Claude or another MCP client, without rebuilding the analysis.
- Export a finished report with APA/MLA/Chicago/BibTeX citations and embedded R source code for use in papers, presentations, or regulatory filings.
Tools this server exposes
| Tool | Does |
|---|---|
| create_analysis | Commission a new statistical analysis by specifying data source (CSV upload, public URL, or OAuth connector), a question, and desired depth tier (Snapshot, JSON |
| list_analyses | Query your private library of commissioned analyses stored in your account. |
| rerun_analysis | Rerun a saved analysis on fresh data without rebuilding it. |
| get_analysis_report | Retrieve a finished analysis report including charts, insights, citations, and source code. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.