Graded, never guessed
Two scores per compound: effectiveness for its purpose, and safety. Cited, or it refuses.
Platform
One operating system for your whole stack.
From the weekly bench to fully-cited dossiers, every feature turns scattered protocols and bloodwork into grounded, cohort-backed signal.
Decision support, not medical adviceAnimations are simplified; real features are more comprehensive
Why it’s different
Not a chatbot. Not a database. Not a forum.
The capabilities a serious stack demands, and who actually delivers them.
| Capability | General AI | Supplement DBs | Forums | Your clinician | SelfAssay |
|---|---|---|---|---|---|
| Grades effectiveness and safety, per compound | Partial | Partial | No | No | Yes |
| Real-world outcomes at scale, not just studies | No | No | Partial | Partial | Yes |
| Reasons over your specific stack | Partial | No | No | Partial | Yes |
| Flags interactions across your whole combination | Partial | Partial | No | Partial | Yes |
| Measures whether it's working for you (n-of-1) | No | No | No | Partial | Yes |
| Every claim cited, or it refuses when thin | No | Partial | No | Partial | Yes |
| Calibrated confidence, not false certainty | No | Partial | No | Yes | Yes |
| Private, member-owned data (export & delete) | No | No | No | Partial | Yes |
Each is good at what it’s for. SelfAssay is the only one built to reason over your whole stack with cited evidence, and refuse when it can’t.
Provenance you can see
Truth is in the provenance.
Real-world reports, peer-reviewed literature and a curated knowledge graph, plus clinical trials, pharmacokinetics, adverse-event and FDA-label signals. Every answer is traceable to its source, with its method and sample size shown.
SelfAssay
Peer-reviewed
studies
114K+
Real-world
reports
181K+
Knowledge graph
relations
186K+
Evidence routed to one answer
SelfAssay
Peer-reviewed
114K+
studies
Real-world
181K+
reports
Knowledge graph
186K+
relations
Biomarkers
Pharmacokinetics
Bioactivity
Adverse events
FDA labels
Clinical trials
181K+
Classified user reports
114K+
Peer-reviewed studies
186K+
Knowledge-graph relations
450+
Compounds graded
How we grade evidence
Rigor you can audit.
Grounding isn’t a tagline; it’s a method. Here’s the standard every answer is held to.
- 01
Triangulated, not single-source
Every signal is cross-checked across classified real-world cohorts, peer-reviewed literature, and the curated knowledge graph; agreement and conflict are both surfaced.
You see corroboration, not just citation. - 02
Every claim carries its receipts
Source, method and sample size shown inline. No naked assertions. If it’s stated, you can trace where it came from.
Follow the trail from claim to original data. - 03
Confidence is calibrated
Strength of evidence is graded; weak signals are labelled weak. Precision is never overstated to sound more certain than the data allows.
You get the truth, with its uncertainty. - 04
“Not enough data” is a valid answer
When grounding fails, SelfAssay says so plainly, and names what evidence would change the answer. It never fills the gap with a guess.
Better to be clear about limits than confidently wrong.
