Skip to content

Whose terminal it is in

The TUI is a guest in somebody's terminal, and the host changes what "export" means. It detects where it is running and delivers accordingly.

The header tells you which one it decided on, so you never have to guess:

scelo tui · qwen2.5:7b-instruct-q4_K_M · ready · ⇩ export → RStudio (click here or ctrl-e)

The four hosts

host detected by what /export does
RStudio RSTUDIO=1 writes flat into the open project, your working directory, as <stem>_analysis.R, <stem>_data.csv and the rest. They appear in the Files pane instantly, and the chat hands you the exact source("<stem>_analysis.R") line. Nothing to open: the project you are in is the export.
VS Code TERM_PROGRAM=vscode, or VSCODE_PID / VSCODE_CWD exports to the usual directory, then runs code -r on the script, notebook and R file, so they open in the window you are already in and whatever extensions you have claim them by file type. Forks are covered: cursor, windsurf, codium and code-insiders are probed when code is absent.
Scelo IDE SCELO_IDE=1 writes flat into the open workspace, SCELO_IDE_WORKSPACE, visible in the file browser immediately, with the .sce one drag away from the Scelo screen.
plain terminal none of the above today's tidy <stem>.scelo-export/ directory.

Flat mode prefixes everything

In RStudio and the Scelo IDE the files land directly in a directory you are already working in, so every generic name gets the dataset stem in front of it: book_data.csv, not data.csv. Dropping a file called data.csv into somebody's project root is a collision waiting to happen.

The scripts are generated to read the prefixed name, so source() and Run All work unmodified. Compare:

book.scelo-export/
  data.csv
  analysis.py
  analysis.ipynb
  analysis.R
  book.xlsx
  book.sce

./                       ← your open project
  book_data.csv
  book_analysis.py
  book_analysis.ipynb
  book_analysis.R
  book.xlsx
  book.sce

The .xlsx and .sce already carry the dataset name, so they are not prefixed twice.

Testing a host locally

Setting the variable is enough, which is also how you check what a colleague will get:

RSTUDIO=1 scelo book.csv          # pretend to be RStudio
TERM_PROGRAM=vscode scelo book.csv
SCELO_IDE=1 SCELO_IDE_WORKSPACE=$PWD scelo book.csv

Why detect at all

Because the right answer genuinely differs. In RStudio, a tidy sub-directory is an obstacle: you would have to leave the project to reach it. On a plain terminal, writing six files into whatever directory somebody happened to be standing in is rude. The same command should do the locally correct thing, and the only way to do that is to know where it is.