NutriStack Research/Edition 1.0·Updated June 1, 2026·4,682 source records

How we built this.

The data model, evidence grading, interaction methodology, and sourcing behind NutriStack Research.

NutriStack is independently owned.
No manufacturer payments. No affiliate revenue. No celebrity protocols.

Conflicts policy · About the team

Stack Score

The exact formula, no black box.

Your Stack Score is a weighted 0–100 composite of safety, evidence, routine quality, and stack load. The source code lives in the app.

Composite weights

Safety
45%
Evidence
25%
Routine
20%
Load
10%
Final range
0–100

Safety carries the most weight. Synergy bonuses are capped, and conflicts or contraindications can cap the final score until they are resolved.

What changes the score

The score now reacts to more than pairwise interaction counts. It uses interaction severity and evidence, ingredient evidence tiers, whether doses and timing are filled in, food-context fit, active stack size, and category redundancy.

Strong evidence ingredients
raise confidence
Missing doses
lower routine quality
Food/timing mismatch
lowers routine quality
Large redundant stacks
lower load clarity
Contraindications
cap the final score

The in-app Stack screen shows the four dimensions and the current score drivers for the active routine.

A note on the in-app combined score

Earlier versions used a base score plus interaction deltas. The current in-app Stack Score uses the four-part weighted model above, so it can respond to evidence quality, schedule completeness, and stack complexity as well as safety.

Interactions

2,504 records across 579 substances. Every one cited.

The interaction database is the heart of the app. Here is exactly what is in it and where every entry came from.

Each interaction row contains

Where the data comes from

Interaction records draw from attached PubMed-indexed studies, reviews, regulatory labels, and curated pharmacology references. A source may support one ingredient, mechanism, or safety concern without testing the exact pair; prescription-pair pages are eligible for search only when the evidence ledger finds a direct human source naming both substances. Citation presence is not the same as clinical review.

Citations include the study text, PMID (for PubMed-indexed studies), and DOI where available. Most rows include the strongest 2–4 studies behind that specific interaction. The full set of source links is browsable at /research/interactions.

Evidence tiers

Five tiers, plain English.

Every claim, supplement profile, and interaction carries a tier. Tiers are not letter grades and not stars. Here is what each one means.

Strong

Multiple high-quality RCTs or meta-analyses.

Effect has been replicated in independent randomized trials or pooled in systematic reviews. The biological mechanism is well-understood. The highest tier we assign.

Moderate

Consistent evidence from quality studies.

Multiple studies point the same direction, often including at least one RCT. Some questions about dose, population, or duration may remain open.

Emerging

Promising but preliminary.

Smaller studies, single trials, or strong mechanistic evidence without large-scale confirmation yet. Flagged clearly so hope is not mistaken for proof.

Limited

Sparse or conflicting evidence.

Only a handful of studies exist, the results disagree, or the study designs were weak. We say so plainly rather than averaging the noise into a confident-sounding score.

Traditional

Long historical use, limited modern science.

Centuries of traditional use, often in Ayurveda or TCM, but minimal randomized evidence in modern populations. Reserved for substances with that specific profile.

Peptides

Peptide coverage is split by regulatory status.

A peptide is not automatically a supplement, and popularity is not evidence. NutriStack separates nutritional peptides, FDA-approved peptide drugs, and research-use products before it considers any recommendation.

Core regulatory sources reviewed for this policy include FDA's compounded bulk-substance safety-risk table, FDA's unapproved GLP-1 warning, DailyMed's Zepbound label, and the 2026 USADA/WADA guide.

Updates

How the library stays current.

The database ships with the app, and every app update can refresh it.

Conflicts policy

No payments. No celebrity protocols.

The conflicts of interest we will not have.

See it in practice

The methodology, working on a real pair.

The worked example on the homepage walks through Atorvastatin + Coenzyme Q10 in the same shape used across the database: what is happening, the mechanism, the recommendation, and the primary literature.

Read the worked example → Browse the full research index

NutriStack is an informational and organizational tool, not a medical service, and not a substitute for professional advice. Always consult a qualified healthcare professional before starting, stopping, or changing any supplement or medication.