The Science Behind Resin Health
La ciencia detras de Resin Health
Every answer grounded in NIH data. Every claim cited. No hallucinations, no guesswork.
Our Data Sources
Five NIH databases, one unified interface
Cinco bases de datos del NIH, una interfaz unificada
We don't generate health information — we organize and present what NIH has already validated.
ODS Fact Sheets
Office of Dietary Supplements
100+ consumer and professional fact sheets covering vitamins, minerals, and botanicals. Available in English and Spanish. Updated quarterly by NIH researchers.
DSLD
Dietary Supplement Label Database
120,000+ supplement products with complete ingredient lists, dosages, and manufacturer data. Searchable by name or barcode.
DailyMed
Drug-Supplement Interaction Data
Drug-supplement interaction data from FDA-approved Structured Product Labels. The same data your pharmacist uses.
OpenFDA
FDA Adverse Events & Recalls
Adverse event reports and supplement recalls. Real-world safety signals from the FDA's post-market surveillance system.
RxNorm
National Library of Medicine
Drug name normalization from the National Library of Medicine. Ensures we match your medications correctly regardless of brand vs. generic name.
How It Works
From your bottle to your safety dashboard
De tu frasco a tu panel de seguridad
Step 1
User searches or scans
Step 2
Match against DSLD
Step 3
Extract ingredients
Step 4
Cross-reference DailyMed
Step 5
Classify severity
Step 6
Plain-language dashboard
AI Chatbot
Retrieval-Augmented Generation — no hallucinations by design
Generacion aumentada por recuperacion — sin alucinaciones por diseno
Our AI chatbot never invents information. It retrieves verified NIH content first, then generates a plain-language answer from only that content.
1
Pre-indexed ODS fact sheets split into searchable chunks — each chunk tagged with its source document and section.
2
User question matched against relevant chunks by keyword similarity — only the most relevant passages are selected.
3
Claude generates an answer ONLY from retrieved NIH context — never from its general training data.
4
Every response includes source citations with links to the original NIH fact sheet for verification.
5
If the retrieved context doesn't contain the answer, the chatbot says: "I don't have enough information from NIH sources to answer that."
Our Commitment
What we do — and what we don't
Lo que hacemos — y lo que no hacemos
What We Don't Do
Diagnose medical conditions
Recommend specific treatments
Make claims beyond NIH data
Store personal health information
Replace your healthcare provider
What We Do
Cite every claim with its source
Link to original NIH fact sheets
Flag interaction severity clearly
Use plain language (6th-8th grade reading level)
Recommend consulting healthcare providers
Transparency
Data Freshness
Frescura de datos
Every data source is versioned, timestamped, and traceable to its origin.
| Source | Records | Last Updated | Update Frequency | Provenance |
|---|---|---|---|---|
| ODS Fact Sheets | 100+ | March 2026 | Quarterly | ods.od.nih.gov |
| DSLD | 120,000+ | March 2026 | Monthly | dsld.od.nih.gov |
| DailyMed | Full database | March 2026 | Weekly | dailymed.nlm.nih.gov |
| OpenFDA | 500,000+ events | March 2026 | Weekly | open.fda.gov |
| RxNorm | 200,000+ terms | March 2026 | Monthly | rxnorm.nlm.nih.gov |