# RMCP MCP server

> Runs R statistical analysis, regression, time series, and machine learning on your data from a natural-language conversation.

- Repository: https://github.com/finite-sample/rmcp
- Page: https://hostedmcp.org/servers/finite-sample/rmcp
- Transport: —
- Stars: 217
- License: MIT

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `RMCP_API_KEY` | no | no |  |
| `RMCP_HTTP_PORT` | no | no |  |
| `RMCP_LOG_LEVEL` | no | no |  |
| `RMCP_R_TIMEOUT` | no | no |  |

## FAQ

### What kinds of statistical analysis can RMCP run?

It handles linear and logistic regression, panel data, instrumental variables, ARIMA forecasting, clustering, decision trees, random forests, and hypothesis tests like t-tests and ANOVA.

### How many R packages does RMCP make available?

RMCP bundles a whitelist of 429 R packages drawn from CRAN task views across more than 19 categories.

### Does RMCP require authentication for remote use?

Yes, remote HTTP deployments require a bearer token, while a localhost server needs no auth.

### How do I install RMCP?

Install it with pip and run rmcp start; Python 3.11+ and R 4.4.0+ are required.