ai.aviado/health MCP server

Find which supplements help with a health condition or goal (e.g., 'sleep', 'anxiety', 'joint pain'). Returns evidence-graded supplement recommendations with dosages.

Find supplement-supplement interactions (765 pairs from clinical evidence) and drug-supplement interactions (1,626 FDA-validated pairs). Returns synergies, antagonisms, absorption conflicts, and timing recommendations.

Find registry-classified supplements in Aviado's current graph that affect a biomarker. Excludes food_substance, prescription_drug, research_compound, research_nootropic, medical_compound, and unknown IDs. Type labels and graph_status (proven or suspected) are internal taxonomy/eligibility metadata, not independent clinical verification. Results are partial and non-exhaustive; pathway_tags are raw graph metadata, not validated mechanisms. Use dose_value with dose_unit; legacy dose_mg is populated only for literal mg rows.

Find which biomarkers are affected by a given supplement. Returns direction (up/down), confidence, dose, and health goal mapping.

Get comprehensive information about a supplement: what it is, what it does, which biomarkers it affects, mechanism of action, and safety contraindications.

Resolve a supplement identity. Use mode='supplement_retail' only for raw retail product titles; it invokes the fail-closed retail front door. Identity mode is the default.

Find the highest-quality brands for a supplement, ranked by ProofMark (Aviado's quality score: third-party testing, certifications, and quality metrics).

Search for a health condition by name with fuzzy matching. Use this to resolve ambiguous condition names before calling get_supplements_for_condition.

Find mixed interventions in Aviado's current graph that affect a biomarker, labeled with the exact intervention_type value supplement, food_substance, prescription_drug, research_compound, research_nootropic, medical_compound, or unknown. graph_status is the exact internal proven/suspected eligibility label, not independent clinical validation. Coverage is partial and non-exhaustive; absence is not evidence of no effect. pathway_tags are raw graph metadata, not validated mechanisms, and dose/pathway details are suppressed for non-supplement and unknown rows. Use dose_value with dose_unit; legacy dose_mg is populated only for literal mg rows.