About · Transparency
About Myto Intelligence
A research-grade, claim-safe lookup engine that answers one question for licensed clinicians: What plants or herbs share mechanism or indication with this medication?
Why this exists
Patients ask their clinicians about plant alternatives constantly. PubMed alone is too noisy. Examine.com is consumer-framed. Traditional herbal references lack drug-context structure. The result: clinicians either dismiss the question or freelance their answer. Both fail the patient.
Myto Intelligence gives clinicians a starting point — mechanism-mapped, evidence-tiered, claim-safe — for the conversation they're already having with patients about botanical alternatives. We don't replace clinical judgment. We surface structured evidence so your judgment has a foundation.
Currently 622 medications across 349 plants, fungi, and ethnobotanicals, mapped through 110 mechanism targets, with compound-level pharmacology for every entry. The dataset is ongoing — new additions are made as primary sources are processed.
How the dataset is built
Each entry is curated from primary sources — published research, systematic reviews, regulatory documents (FDA, German Commission E, WHO), and peer-reviewed reference texts. We follow a structured evidence ladder (A through G):
- Tier A: Systematic review or meta-analysis of human studies
- Tier B: Randomized controlled human trial
- Tier C: Observational human study
- Tier D: Animal study
- Tier E: In vitro or mechanistic study
- Tier F: Traditional or ethnobotanical record
- Tier G: Hypothesis or in-silico only
Tiers F and G are never presented as proof of clinical effect. Citations are categorized by verification status — see the Citations page for full transparency on PMID-confirmed vs. PMID-pending entries.
Privacy & data architecture
No PHI by design
Myto Intelligence is a clinician reference tool — like UpToDate or Epocrates — not a PHI-handling platform. We are not a HIPAA covered entity and we do not act as a business associate. Patient information has no place on this platform; clinicians are responsible for de-identifying their inquiries before submission.
Our system is designed to be structurally incapable of holding patient identifying information. Conversations submitted to the AI agent (when live) pass through automated PHI-detection that scans for and redacts common identifying patterns (names, dates of birth, MRN-like numbers, addresses) before any storage or external API call. This is a backstop — not a substitute for the clinician's no-PHI discipline.
What we store
- Account information: email, name, NPI, state of practice, specialty, intended use
- Subscription state: tier, access dates
- AI conversation history: stored per-clinician for personal reference; subject to PHI-scrubbing on input and output; deletable on demand
- Audit metadata: which clinician accessed which page when (no content)
What we do not store
- Patient names, dates of birth, addresses, MRNs, or any identifying information
- Patient case details with sufficient identifying specificity
- Information shared with us by third parties about identified individuals
Where data lives
User accounts and conversation history are stored in Supabase (PostgreSQL with encryption at rest, US data center). AI inference runs through Anthropic's Claude API. The application is hosted on Vercel. No data crosses to external advertising or analytics networks.
What this is — and is not
It is
- An educational reference tool for licensed clinicians
- A research-stage, invite-only preview
- A starting point for the conversation between a clinician and a patient about plant alternatives
- Continuously updated as primary sources are processed
It is not
- A clinical decision-support tool
- A substitute for clinical judgment
- A diagnosis, treatment, or prescription tool
- A FDA-regulated medical device
- A HIPAA-regulated PHI handler
- A complete reference — coverage gaps exist and are documented
Public launch is gated on completion of citation verification (see Citations) and FDA / FTC-experienced legal review of the result-page template. Until then, this is invite-only research preview.
Feedback & contact
The most valuable thing you can do is tell us what's missing or wrong. The dataset gets stronger with every correction or coverage-gap report. Reach out:
- Founder & Operator: Daniel J. Klimczak — mytointel@mytointelligence.com
- For dataset corrections, citations, additions: same email; subject line "Myto Intelligence — [topic]"