Claims should map to sources. Risk should map to evidence quality.
The entire point of StackMax is to make research legible without flattening nuance. Strong claims require strong sources. Weak evidence stays visibly weak.
How evidence should flow through the product
Each compound profile starts with structured research notes: mechanism, plausible benefits, known risks, regulatory context, dosing conventions, and cited studies. Those studies should never sit as decoration. They should directly support the specific claim shown to the user.
Stack analysis then reasons across compounds rather than treating each one in isolation. The job is to surface overlapping pathways, contraindications, cumulative risk signals, and obvious blind spots in the literature.
How confidence is represented
StackMax uses evidence scoring to distinguish strong human evidence from weak preclinical signals. A claim backed by one in-vitro paper should not look the same as a claim backed by multiple randomized trials.
Where the literature is thin, the UI should say so plainly. Uncertainty is part of the product, not something to hide.
What the current demo is and is not
The current build demonstrates the product direction with curated mock data. It shows the reporting style, risk framing, and interaction model we want, but it is not yet a live ingestion pipeline from external databases.
That is deliberate. The first step is getting the decision surface right. The second step is scaling the dataset and automation behind it.