How to vibe-code an experiment¶

You can let an AI coding agent (Claude, Gemini, etc.) write the experiment for you. Open the agent inside the goxpyriment folder and describe your experiment (stimuli, design, etc.) in plain language. Asking it to add a new experiment to the examples folder leads it to read the existing examples for context. Recommendation: save your prompt in a description.md file.

Example: reproducing an experiment from a paper¶

Create a folder for the new experiment, copy the paper into it, and start the agent from the repository root:

cd goxpyriment
mkdir -p examples/Finger-Tracking
cp Dotan_Cognition.pdf examples/Finger-Tracking
claude

Switch the agent to plan mode, so that it proposes a plan before writing any code:

❯ /plan
  ⎿ Enabled plan mode

Then describe what you want:

❯ Can you read the method section of the pdf file located in
  examples/Finger-Tracking, then create a plan to reproduce their
  finger-tracking experiment where the trajectory of the finger on a touch
  screen, or of the mouse, is measured while the participants must attain a
  target on a horizontal numerical line at the top of the screen? You can ask
  me questions if anything is unclear from the method section of the paper.

The Finger-Tracking example in the gallery implements this experiment.