Open a Pharo Playground and evaluate:
"Option 1: Open empty dashboard, create everything from buttons"
LLMDashboardPresenter open.
"Option 2: Open dashboard with a pre-built pipeline"
LLMDashboardPresenter openOnPipeline: (LLMPipeline new
addStep: (LLMPromptStep template: 'Translate to {language}: {text}');
addStep: (LLMModelStep model: LLMOpenAIModel gpt4o);
addStep: (LLMParserStep parser: LLMStringOutputParser new);
name: 'Translator';
yourself).
Then use the toolbar buttons: New Pipeline to create, Run to execute, Re-Run to repeat.