Skip to content

Diligence-grade source list

Our Data Sources

Every dataset behind a Avelor verdict, with the upstream organization, refresh cadence, and current ingestion status. See methodology for how these sources combine into a score; see about our data for the longer-form explanation.

How sources become consumer-visible answers

$Avelor runs an AI-assisted clinical review engine over every primary source listed below. Auto-extracted candidate interactions are graded against published mechanism + evidence weight + severity tiers before they enter the consumer corpus. We never publish raw model output as fact.

openFDA Adverse Events

U.S. Food and Drug Administration
Live

FDA Adverse Event Reporting System (CAERS) supplement-tagged events, used for safety-signal flagging on every scan.

Daily refresh772 CAERS events ingestedSource ↗

FDA Recalls (Class I, II, III)

U.S. Food and Drug Administration
Live

Active and historical FDA enforcement actions filtered to dietary supplements.

Daily refresh2,354 supplement recalls ingestedSource ↗

FDA Tainted Products

U.S. Food and Drug Administration
Live

FDA list of supplements found to contain undisclosed pharmaceutical ingredients (e.g. weight-loss with sibutramine, sex-enhancement with sildenafil).

Weekly review2,061 tainted-product recordsSource ↗

NIH Office of Dietary Supplements (ODS) Fact Sheets

National Institutes of Health
Live

Per-ingredient evidence summaries with consumer + health-professional variants. Backbone of the dense-search corpus for general supplement questions.

Synced on NIH publication87 fact sheets ingested into RAGSource ↗

NCCIH Clinical Digests

National Center for Complementary and Integrative Health
Live

Practitioner-grade summaries for complementary and integrative supplements, used for evidence-grade questions.

Synced on NCCIH publication47 clinical digests ingestedSource ↗

PubMed E-utilities

National Library of Medicine
Live

Citation lookup for every published claim. Avelor only cites PMIDs the system can verify against the live PubMed API.

On-demand per-ingredient1.39M studies indexed across the corpusSource ↗

USDA FoodData Central

U.S. Department of Agriculture
Live

Nutrient composition reference for whole-food alternatives surfaced alongside supplement recommendations.

Quarterly USDA releasesSource ↗

NIH Dietary Supplement Label Database (DSLD)

National Institutes of Health
Live

Source of truth for the product catalog: labels, ingredients, manufacturers. We layer scoring + safety signals on top.

Continuous (NIH-driven)178,763 products + 1,995 brands ingestedSource ↗

DailyMed (FDA Structured Product Labels)

U.S. National Library of Medicine
Live

FDA-managed structured product label database for prescription and OTC drugs. Used to verify prescribing-information facts before any drug-supplement interaction assertion.

Continuous (NLM-driven)Source ↗

supp.ai-derived Candidate Queue

Allen Institute for AI (research-only reference)
In review

Auto-extracted supplement-drug interaction candidates from biomedical literature. Each candidate must pass our AI-assisted clinical review engine (pharmacist-style structured grading on mechanism, evidence weight, and severity) before promotion into the published interactions corpus consumers see.

Periodic re-extraction from primary literature27,578 candidates awaiting clinical adjudicationSource ↗

Health Canada / EFSA / TGA / WADA

International regulatory bodies
Live

Cross-jurisdictional safety + sport-banned-substance signals layered on top of FDA-only data for international travelers and competitive athletes.

Quarterly reviewSource ↗

Clinical advisory board

Avelor is recruiting a clinical advisory board (one practicing pharmacist, one functional-medicine MD, one registered dietitian) to review interaction-severity classifications quarterly. Roster will be published here once formalized. Review in progress.

Snapshot date: 2026-09-25. Counts refresh on deploy. Discrepancies? Email hello@askavelor.com.