AI lab analysis · built for functional medicine

Every patient's labs, analyzed in seconds. Reviewed and signed by you.

Zenlo Labs reads a full panel across 102 biomarkers and 15 clinical patterns (8 validated on 4,018 NHANES · 7 implemented), drafts a structured analysis with HOMA-IR and biological age, and hands it to you to edit, sign and share. The AI does the synthesis — the clinical decision stays yours.

Validated on 4,018 NHANES patients medRxiv preprint HIPAA-aware · Anthropic BAA (ZDR)
Health composite
82 / 100
Optimal markers37
Attention markers4
Critical findings1
Biological age
47 yrs
+5 vs chronological
HOMA-IR
3.1 High
The doctor dashboard

Open the app and see what needs you.

Your first screen isn't a welcome banner — it's the state of your practice: critical findings, what's new, and who's waiting on a review.

Good morning, Dr. Reyes
All patients + New patient
Critical findings
1
Recent labs
14
Pending review
3
Action items
6
Recent activity
Jane D. — critical finding flagged (insulin resistance)2h
Marcus T. — new panel uploaded, 41 markers5h
Alan W. — report signed & locked1d
Sofia M. — borderline CVD pattern detected1d
Needs attention
Jane D. · 42F
Insulin resistance · awaiting review
Sofia M. · 45F
CVD risk · borderline
⌕ Search patients
All Critical Attention Healthy
128 patients
JD
Jane Doe
42F · 38 MARKERS
28 MayCritical
SM
Sofia Marin
45F · 30 MARKERS
27 MayAttention
MT
Marcus Tan
52M · 41 MARKERS
26 MayProcessing
AW
Alan Wu
39M · 35 MARKERS
24 MayHealthy
Your patients

The whole panel population, one search away.

Find anyone in seconds, filter by severity to triage your day, and open straight into their latest analysis. Built to stay fast even with thousands of patients.

  • Severity filters — see only critical or borderline patients when you're triaging.
  • Status at a glance — critical, attention, healthy, or processing on every row.
  • Trends over time — every repeat panel tracked per patient.
Inside a report

A structured analysis, not a wall of numbers.

Each panel becomes a composite score, the patterns that matter, the findings that need action, and the trends over time — ready for you to edit, note, and sign.

JDJane Doe · 42F
PANEL · 38 BIOMARKERS · 28 MAY
Add noteSign & lock
Health composite
82 / 100
Optimal37
Attention4
Critical1
Detected patterns
Insulin resistanceCritical
InflammationAttention
Liver riskIn range
Critical finding
HOMA-IR3.1
Clinical threshold > 2.5High
HbA1c · trend5.9%
5 PANELS · TRENDING UP
Trigger
Labs uploaded for a patient
38 BIOMARKERS PARSED · RANGES MAPPED
Engine
Three-tier pattern detection
PATHOGNOMONIC · CONFIRMATORY · SUPPORTIVE — 15 CLINICAL PATTERNS (8 VALIDATED · 7 IMPLEMENTED)
Compute
HOMA-IR + biological age
DERIVED METRICS · CLINICAL THRESHOLDS
Output
Draft report → your dashboard
EDIT · ADD NOTES · SIGN & LOCK
What's under the hood

A clinical engine, not a chatbot.

Built around a defensible detection framework and validated against population data.

102

Biomarkers

A full panel mapped to age- and sex-specific reference ranges.

15

Clinical patterns

15 clinical patterns (8 validated on 4,018 NHANES · 7 implemented).

Three-tier detection framework

Each finding is graded by evidence — pathognomonic, confirmatory, supportive — with compound rules and N-of-M cluster thresholds, so a flag means something.

HOMA-IR

Insulin-resistance index computed automatically from glucose & insulin.

Biological age

Levine PhenoAge (2018) when nine required biomarkers are available.

Notes · Sign & lock

Inline editor, your wording, a tamper-evident signed record.

One-click PDF

Clean, print-ready report to share with the patient.

Privacy by design

Anthropic BAA with zero data retention; HIPAA-aware.

How it works

AI drafts. You decide.

Three steps from a raw panel to a signed report — and you stay in the loop on every one.

01

Upload the labs

PDF upload or manual entry. Values are extracted and mapped to reference ranges.

02

AI drafts the analysis

The engine detects patterns, computes HOMA-IR and biological age, and writes a structured draft to your dashboard.

03

You review & sign

Edit the wording, add your notes, then sign and lock. The clinical judgment is yours.

Transparency

Model cards, biomarkers, and security — in the open.

CHAI-format pattern cards, a live biomarker registry, and a dated security self-assessment. 102 biomarkers in product copy; 15 clinical patterns with published NHANES audits for Insulin Resistance and Metabolic Syndrome.

15 clinical patterns. 8 benchmarked across 5 AI models on 4,018 NHANES adults (F1 up to 0.963). Per-pattern NHANES audits published for 2 (Insulin Resistance, Metabolic Syndrome); the rest are structural CHAI cards, validation in progress.

→ All validation audits

Evidence

Validated, published, and open about its limits.

The detection engine was benchmarked across five AI models and four providers on a real population cohort — 15 clinical patterns (8 validated on 4,018 NHANES · 7 implemented). We publish our methodology rather than asking you to take it on faith.

0.963
Best-model F1
4,018
NHANES patients
5 / 4
Models / providers
102
Biomarkers
Get started

See Zenlo Labs on your own panels.

Bring a few de-identified reports and we'll walk through the analysis live — patterns, HOMA-IR, biological age, and the review-and-sign flow.

Open Zenlo Labs → Request a demo