Clusy now runs R natively

Pick R when you start a notebook and the agent writes the R, runs it, checks the plots, and gives you the finished notebook. A real R session, in the same workspace as your Python one.

Product17 September 20262 min read

Until today, Clusy spoke Python. If your work lived in R, you were asked to translate it, or to keep a separate R install for the parts that would not move. That is over.

Clusy now runs R natively. A real R session, in the same workspace as your Python one, with the same agent. Pick R when you start a notebook and describe what you want. The agent writes the code, runs it, checks the plots, and gives you the finished notebook.

Survival analysis, mixed models, anything on CRAN. The things you kept R around for.

One notebook, two languages. Each cell keeps its own.

Pick R. Then just ask.

Choose R on the home page or in the first message of a new notebook. Clusy remembers the choice for next time. Each cell carries its own language, with R highlighting, completion and diffs, so one project can hold Python and R side by side without either one pretending to be the other.

The agent knows which language a cell is in. When you ask for a change, it edits R as R.

One run, start to finish

The clip below is a single run, sped up, with nothing edited out. The ask was a survival analysis on the NCCTG lung cohort that ships with the survival package: Kaplan-Meier curves by sex and by ECOG performance status with numbers at risk and log-rank tests, a Cox model with a hazard-ratio forest plot, a proportional hazards check with cox.zph and Schoenfeld residuals, and a summary a clinician could read.

Seven minutes on a CPU sandbox, played back fast. The forest plot's labels clipped on the first pass; the agent saw it and fixed the margins before writing the summary.

Two things in that run are worth pointing at. The agent read its own plots and corrected a layout problem without being asked. And the numbers it reported are the ones on screen: a hazard ratio of 0.55 for female sex, 1.67 per ECOG point, and a global cox.zph p of 0.31, so the proportional hazards assumption held.

Your session stays put

The R session is saved with the notebook. Branch it, close the tab, come back next week, and your objects are still there. Packages you install go into the branch's own library and stay installed. When you are done, export as .ipynb, R Markdown or Quarto.

It is already on your account. Open a notebook, pick R, and give it something real.

If part of your R work does not fit yet, tell us what it is. It decides what we build next.