ABOUT · LONGEVITY ATLAS

We built the evidence layer we wished existed.

Longevity information is abundant. Decision-grade evidence is not.

62Interventions graded
18,844+Trial records indexed
WeeklyEvidence refresh
v1.2Current methodology

Last updated · June 2026 · sourced from ClinicalTrials.gov and PubMed

§ 01 — Why this exists

Honest information is the only kind that is useful.

Too many decisions in the longevity economy are made on marketing copy, not data. Brands cherry-pick the most favorable trial; mouse studies get extrapolated into human claims; the actual clinical evidence — thousands of registered trials, meta-analyses, null results — sits in databases most buyers never open.

Longevity Atlas is an AI-native evidence intelligence company building the decision layer for the longevity economy. We map what the evidence actually says, including the weak results and the failures, and we treat East and West by the same standard: decades of rigorous research on Chinese tonic herbs and adaptogens lives mostly in Mandarin-language literature and regional registries. We index both sides and grade them identically.

§ 02 — Who is accountable

We stand behind every grade we publish.

phycyber
Research & intelligence studio · Longevity economy

Longevity Atlas is built and operated by phycyber. We develop our own tooling to index trial registries and scientific literature at scale, and we are accountable for every confidence score on this site. A real person answers research@phycyber.ai.

Evidence Governance & Review Network

Pharmacology · Nutrition science · Regulatory review

Independent specialists benchmark the system, review higher-risk edge cases and advise on methodology — they do not manually author every Intel File. Today more reports pass through human review; as the engine matures, automation expands and human involvement concentrates on governance and exceptions. Named reviewers are listed here only once their credentials are public — we do not pad the team with stock photos or titles we cannot verify.

§ 03 — How we work

How the work is organized.

The evidence engine produces the indexing, screening, grading and citations; people define the methodology, audit the output and handle high-risk exceptions. The methodology explains the scoring rules — this is how the work is run, and where responsibility sits at each step.

01
INDEX
  • Primary research and trial registries
02
SCREEN
  • Human relevance, study design and endpoints
03
GRADE
  • Quality, uncertainty and safety — automated, with human escalation on risk
04
DECIDE
  • A sourced, decision-ready Intel File
§ 04 — What we believe

Three commitments we will not trade away.

Null results are data.

If a trial found no effect, we report it. Cherry-picking is how weak products borrow credibility; we show the full picture, including the failures.

Every grade is traceable.

Each confidence score traces back to the primary sources behind it — registries, RCT publications, meta-analyses. If you can see how we arrived at a number, you can challenge it.

We state the boundaries.

An honest grade says what the evidence does not show. Maturity is not effect size, and a high score is not a medical recommendation.

§ 05 — Conflict of interest

How we make money, and what it does not buy.

Revenue sourceInfluence on public scoresDisclosure
Intel Files ($799+)NoCommissioned research is separated from public grading.
Supplier relationshipsNo paid rankingListing is never sold; evidence rules remain unchanged.

If you believe a conflict of interest has affected our research, tell us: research@phycyber.ai.

WORK WITH US

Submit an ingredient. Get a citable Intel File in 7 days.

For brands and formulation teams who need primary-source evidence intelligence before betting on an ingredient. Starting at $799.

Submit an ingredient →Our methodology