Help

AI Recommendation Generation

Updated February 22, 2026

This content is for informational purposes only and is not a substitute for professional advice.

Unfair generates recommendation bundles from your onboarding profile and local supplement catalog data.

Inputs used by the engine

The engine scores candidates from goal, stimulant sensitivity, contraindication state, current supplement list, consistency obstacles, and cycle preference.

These values come from onboarding and preference state, then are normalized into a deterministic input hash.

Output produced

A bundle includes recommended stacks, recommended supplements, excluded supplements, rationale snippets, and confidence scores.

The onboarding flow persists this bundle into your plan state for You tab rendering.

Safety-oriented exclusion behavior

High-risk supplement classes are filtered when contraindications are present.

Stimulant-heavy options are reduced or excluded based on your stimulant sensitivity setting.

When generation runs

Current public flow runs recommendation generation during onboarding and then reuses that stored result until profile state is refreshed through onboarding-driven updates.

Use Recommendation Signals to understand how each input can change ranking.