Pipeline prompt config
Select a pipeline, then edit prompts and call parameters. Prompts are stored per pipeline in the database. Use Export to download prompts as JSON and Import to load prompts from a JSON file.
Placeholder tags
Use these in prompts; they are replaced at runtime. Each tag’s source and content are described below.
Under every prompt textarea, Uses: lists the {…} placeholders detected in that field’s text (updates as you type).
A tag is shown in red when it is not yet produced by any earlier step in the pipeline (earlier stages and earlier calls within the same stage, including Python steps).
Each stage heading shows Produces from the pipeline template JSON and all Uses aggregated across that stage’s prompt fields (stages without LLM prompts still appear so you can see what they produce).
Prompts
Input material tags
Labels describing types of input materials expected by this pipeline (e.g. blood test, microbiome report). These are stored per pipeline and can be used to filter or label uploaded materials.
Call parameters