SUPPDB Supplements & Nootropics Dataset
§ 00 · Real Labels · Normalized Dosages · PubChem Chemistry

Real Supplement
Labels,
normalized.

Abstract

SuppDB leads with its Safety & Interaction Layer — 8,300+ supplement × drug interactions, per-product % of the tolerable upper limit, condition contraindications and WADA flags — on top of a structured catalog of 18,000+ real supplement & nootropic products from 2,000+ brands, built exclusively from the public NIH Dietary Supplement Label Database (DSLD). Every active ingredient is normalized to milligrams, proprietary blends are flagged where the dose is hidden, and each compound is enriched with NIH PubChem chemistry — with a source label URL on every record so any fact can be re-verified. A separate Safety layer builds on the normalized doses: 8,300+ supplement–drug interactions, per-product percentage of the tolerable upper limit, condition contraindications, and WADA doping flags — every interaction cited to its public-domain source.

Get the Dataset ↗ ↓ Sample Explorer 115,000+ rows

One-time purchase · CSV / SQLite / JSON delivered by email within 24 h · Public-domain sources

18,000+
Real Supplement Products · 2,000+ Brands
115,000+
Active-Ingredient Records (dose normalized to mg)
Coverage: UPC 86% · Servings 85% · Proprietary-blend flags on 40,000+ ingredient rows · PubChem chemistry on 4,000+ compounds · Source URL on every record. Honest field coverage published up front.
§ 01 · Headline Metrics · Public-Domain Sources
NIH DSLD + PubChem
NIH DSLD
Public-Domain Label Source

Built entirely from the NIH Dietary Supplement Label Database — real brands, UPCs, forms, servings, and ingredient panels. No scraping, no synthetic rows; every record links back to its DSLD label id.

mg Norm.
Dosage Normalization + Blend Flags

Every dose converted to milligrams — mcg, g, and substance-specific IU handled correctly. Undisclosed proprietary-blend doses are flagged, not silently dropped.

PubChem
Chemistry + InChIKey Canonicalization

CID, molecular formula, weight, SMILES, and InChIKey per compound. Rows sharing an InChIKey are the same molecule under different names (Vitamin C = Ascorbic Acid) — canonicalize deterministically.

§ 02 · Live Ingredient Explorer Preview
Free Sample Snapshot
Brand Product Form Ingredient Dose / serving Category PubChem
Thorne Magnesium Bisglycinate Powder Magnesium 200 mg mineral CID 5462224
Nordic Naturals Ultimate Omega 2X Softgel Omega-3 2,150 mg fatty acid CID 56842239
NOW Sports Creatine Monohydrate Powder Creatine 5,000 mg ergogenic CID 586
Life Extension Neuro-Mag Magnesium L-Threonate Capsule Magnesium 144 mg mineral CID 5462224
Bulletproof Sleep Mode Softgel Melatonin 0.3 mg hormone CID 896
Thorne Curcumin Phytosome 500 mg Capsule Curcumin 500 mg botanical CID 969516
NutraBio Ashwagandha KSM-66 600 mg Capsule Ashwagandha 600 mg botanical — (botanical)
Nature's Way Alive! Whole Food Energizer Tablet Proprietary Blend undisclosed (flagged) blend —

Real rows from the free sample. Note the last row: a proprietary blend whose per-ingredient dose is undisclosed on the label — flagged rather than dropped.

Want to dig deeper? Try the interactive Sample Explorer on 🤗 Hugging Face → or download the free 300-product sample dataset.

Browse every page: 514 ingredient monographs · 300 product labels

§ 03 · Three Ways to Get the Data
Safety · Catalog · Custom
Tier 1 · Safety & Interaction Layer

Safety & Interaction Layer

Delivered by email within 24 h · interactions · upper-limit % · contraindications

$199
  • 8,300+ supplement × drug / supplement interactions — a hand-verified core plus entries auto-extracted from FDA drug labels and NIH ODS fact sheets, each evidence-graded with its source snippet so you can filter to the vetted set (NIH ODS, DailyMed, MedlinePlus).
  • Per-product % of the Tolerable Upper Limit with over-limit flags — computed from mg-normalized doses against NIH DRI limits; flags the 1,300+ products that exceed a safe ceiling.
  • Condition contraindications (pregnancy, anticoagulation, renal impairment, pre-surgery, and more).
  • WADA doping flags for banned/monitored compounds, plus a dose-disclosure transparency score per product.
  • Relational SQLite + CSV + JSON, self-joinable via a slim compound index — drops straight into pharmacy, telehealth, and stack-tracker apps.
  • Delivered by email within 24 h of payment · one-time purchase · public-domain sources.
Buy Now · Safety Layer · $199 →
Delivered by email within 24 h · secure Stripe checkout · Apple Pay / Google Pay
Not medical advice — label facts and cited interaction references; verify with a clinician.
Tier 2 · Catalog only

Catalog only — Full Snapshot

One-time download · 18,000+ products · Snapshot 2026.09

$49
  • 18,000+ real supplement products from 2,000+ brands, built entirely from public-domain NIH DSLD labels.
  • Relational SQLite + CSV + JSON: normalized tables (`supplement_products`, `active_compounds`, `product_compounds`, `compound_synonyms`, `data_sources`), ready for AI health apps, dose comparison, and B2B integration.
  • Normalized mg dosages across 115,000+ active-ingredient records, with 40,000+ proprietary-blend flags for undisclosed doses.
  • NIH PubChem chemistry: CID, molecular formula, weight, SMILES, and InChIKey on 4,000+ compounds — plus InChIKey canonicalization of name variants.
  • Full provenance: every record carries its DSLD label id and source URL — traceable and re-verifiable.
  • Snapshot 2026.09 · one-time purchase, no recurring fees. Public-domain sources (NIH DSLD + PubChem).
Buy Now · Full Snapshot · $49 →
Secure card checkout via Stripe · Apple Pay / Google Pay · download link delivered by email after payment
Need a company invoice or crypto payment? use the contact form (suppdb@dataengineered.io)

Custom work (subsets, recurring feeds, schema mapping): from $99 — contact the desk

§ FAQ · Frequently Asked Questions

Technical Supplement & Nootropic Dataset FAQ

How are proprietary blends handled in the SuppDB dataset?

Every ingredient record features an explicit boolean flag (is_proprietary_blend). When a product lists a proprietary blend, the total blend mass is recorded, and constituent ingredients are linked under the shared blend hierarchy without fabricating unverified individual dosages.

Can I get NIH PubChem CIDs and InChIKeys with every supplement ingredient?

Yes! We cross-reference every normalized active botanical and chemical compound against NIH PubChem, enriching records with exact PubChem CIDs, molecular weights, canonical SMILES, and unique InChIKeys.

What format is the dataset delivered in when I purchase the snapshot?

The full dataset snapshot ($49) is delivered by email within 24 h as a clean, relational SQLite database (suppdb.sqlite) alongside normalized CSV tables (products.csv, ingredients.csv, and brands.csv) and JSON dumps.

Can I download a free sample before licensing the full 18,000+ product dataset?

Absolutely. You can download our verified 2,250+ row open sample CSV (samples/suppdb_sample.csv) directly from our landing page or GitHub repository under a Creative Commons license.

§ 04 · Direct Engineering & Dataset Support

Have Questions or Need Custom Data?

Whether you need custom schema formats, API access, or enterprise invoicing, reach out directly to our desk at suppdb@dataengineered.io.

SuppDB is factual reference data compiled from public NIH DSLD labels — not medical advice. Always verify against the current physical product label.