A pair of eyes in a single bar of warm light

Technology

We did not build another test. We built the thing that reads them.

An engine that reads every lab result a person already has, together and over time. It finds what is building up, names the few tests worth doing next with a reason for each, and a registered pathologist signs every report.

17 thousand
Diseasesin our own curated medical knowledge base
50 million+
Literature and linksbuilt on public NIH and JAMA trial data
100 thousand+
Combinationschecked from one report

Why not just a chatbot

Labs have the machines. Nothing reads the results together.

A chatbot predicts the next word. Medicine is maths and evidence. Here code does the maths, rules connect the results, and a pathologist signs.

Your reportPredict the next wordA chatbot aloneCode calculatesRules connectEvidence citesA pathologist signsThe Mutanex engine
  • Two units, so it says "improved"mmol/L vs mg/dLConverted by code, every time
  • Plausible, so it passesTSH 5.6, read as 56Rejected: not in the source
  • One cut-off for everyoneWaist 90 cmThe cut-off for this person, by rule

The engines

Five engines. One shared model of you.

Your case passes through all five, in order. Each adds what it finds to the same digital twin, and the twin grows with every test and every retest.

Over the years

What it does for longevity.

Given everything we know about you: what is the most useful thing to do next, what will change, and when?

Runs on the five engines above. Same machinery, pointed at the years ahead instead of at one disease.

Which test is worth running?

A test earns its place only if a result would change the decision.

Woman, 36, small red cells, TSAT 10%Sample patient, fictional data

≤ 0.20Look for another cause≥ 0.85Treat, and find the sourceIron deficiency 0.75
FerritinEssential
  • LowAbove 0.85Treat, and find the source
  • NormalBelow 0.20Look for another cause
Either result changes the decision
Vitamin DNot recommended
  • Any resultNo changeSame decision either way
Printed on the report, with this reason

Seven decisions, in order

  1. 01KnowYour history and your goal
  2. 02MeasureYour risks, and what is measured
  3. 03TestOnly tests that change a decision
  4. 04RecommendOnly actions backed by evidence
  5. 05PrioritiseOne score per action
  6. 06MonitorEarly markers, real change only
  7. 07ReassessRetest when change could show
It will never say
  • Your biological age
  • Years younger
  • Add years to your life
  • Reverses ageing
  • You have this disease
  • Start or stop this medicine

Calculated, not guessed

Ask a chatbot twice. Get different words.

Ask a chatbot the same question five times and the answer changes. Here, code does the maths and the AI only writes the words. Run both and see which numbers hold.

The same reports, every run (sample patient)
  • ALT 88 to 58 U/L
  • Haemoglobin 11.4 to 12.9 g/dL
  • Ferritin 9 to 28 ng/mL
  • Man, 46

A general AI chatbot (illustration)

A chatbot, asked the same question five times

No history, generic ranges, no one signs it----

Mutanex (sample)

AI writes the words. It never does the numbers

  1. Code does the mathsFIB-4 = (Age × AST) ÷ (Plt × √ALT)
  2. Clinical knowledge connects itIf vegetarian, B12 and iron read differently
  3. AI writes the wordsPlain language only. It never makes the numbers
  4. A pathologist signsReviewed, signed, released

ALT 88 to 58: outside the reference change value, a real fall. FIB-4 0.97: low. Haemoglobin 11.4 to 12.9: a real rise.

Every line cited. Signed by the lab's pathologist (sample)----
Run log
Chatbot
    Engine

      The knowledge underneath

      What medicine knows, organised so code can use it.

      • 01GuidelinesClinical practice guidelines, firstThe threshold behind every flag is a published guideline threshold, with its version.
      • 02ReviewsSystematic reviewsMany trials pooled, graded and dated in one evidence table.
      • 03CohortsLarge cohorts, matched to the personA cohort that matches the person's ancestry, sex and age counts most; others count less.
      • 04ConsensusExpert consensusWhere the trials have not been run, named as such, never dressed up as evidence.
      • 05Text-minedLiterature, mined by machineUseful for finding a lead. Never the source of a patient-facing statement.

      ✓Aligned with published guidelines from WHO, KDIGO, ADA, AASLD, ESC/EAS, the Endocrine Society, NICE and ACC/AHA.

      Longevity first. Precision medicine next, on the same engine.

      Precision medicine comes next: the same reading, supporting your doctor's diagnosis, after two to three years of data inside the labs and hospitals that already run Mutanex.