Apple Watch
58ms
HRV · 7-day avg
“Why did this drop?”
Pre-seed · raising 2026
Health, further.
Further Health turns the fragmented health data you already have into one private, personalized second opinion.
Decision-support, not diagnosis. Informational only — not medical advice.
A personal note
I built this because the system couldn't see me whole.
“I saw eight specialists. None of them could figure out what was going on, until we built our own.
— Zade Kal · Co-founder
Eight specialists, one patient
- Hematologist
- Cardiologist
- Pulmonologist
- Psychiatrist
- Neurologist
- Primary Care
- ENT
- Endocrinologist
38
Blood labs
4
MRIs
2
X-rays
No specialist could connect the signals — until we built the system that could.
The problem · fragmentation
We have more health data than ever, scattered across sources, yet none of it is actionable.
Oura
72
Sleep score
“Is this normal for me?”
Quest / LabCorp
14ng/mL
Ferritin · low
“Should I worry?”
MyChart / Epic
4.5mIU/L
TSH · trending up
“Has it been climbing?”
23andMe
CYP2C19
Pharmacogenomics
“Does this affect my Rx?”
Is my HRV crashing because my ferritin dropped?
The data exists. The answer doesn't and the user is the integration layer.
The problem · trust
And the data they do hand over keeps leaking.
81%
of Americans say the risks of company data collection outweigh the benefits.
Pew Research
71%
are more worried about data privacy than they were a few years ago.
Pew Research
6.9M
genetic profiles exposed in the 2023 23andMe breach, then sold in bankruptcy.
SEC filings · 2024
BetterHelp · $7.8M FTC fineFlo · FTC settlementGoodRx · $1.5M FTC fine23andMe · 6.9M users · bankrupt
Meanwhile a 15-minute primary-care visit can't see wearable data, can't read trends across panels, and judges every value against a population range — not the patient's own baseline.
Our solution
Every health marker, unified into one private profile.
The only platform that ingests every major class of consumer health data — and synthesizes across them. Competitors specialize in one slice. Further owns the whole stack and runs the analysis primarily on the user's own machine.
The core loop
- Connect
- Understand
- Act
- Track
- Improve
A second-opinion engine — not another dashboard.
Ranked hypotheses with confidence levels, supporting evidence drawn from the user's own data, and clear validation paths. Decision-support, not diagnosis.
Inputs
- WearablesWHOOP · Oura · Apple Watch
- LabsQuest · LabCorp · MyChart
- Genetics23andMe · AncestryDNA · TellMeGen
- RecordsMyChart · CCDA · FHIR
- CalendarGoogle · Apple · Microsoft
Further engine
- Cross-domain patterns
- Polygenic risk in context
- Drug-gene interactions
Product · the “aha” moment
3–5 non-obvious insights, minutes after connecting your data.
further.health/me
Tier 1 · localToday, Alex — 1 thing to look at.
Cross-signal · Apple Watch + Quest · confidence 0.74
Your HRV suppression correlates with low ferritin over the last 90 days. Worth a follow-up panel.
Vitals · last 90 days
Ferritin
14ng/mL
↓ 38% · flag
HRV
58ms
↓ 11% · watch
TSH
4.5mIU/L
↑ 22% · watch
LDL-C
94mg/dL
— · stable
Top hypothesis · Bayesian state · gathering evidence
H-001 · Iron-deficient autonomic stress
conf 0.74Declining ferritin (90d) → reduced O₂ delivery → HRV suppression and rising RHR. Validate with CBC + reticulocytes; consider 100mg ferrous sulfate trial.
What no clinician and no other AI can do today.
- Specialists average ~15 minutes per visit, stay inside their specialty, and rarely review years of historical data.
- General-purpose AI hits context-window limits long before it can cross-analyze a decade of HRV, labs, symptoms, and sleep.
- Further connects HRV ↔ symptom dates ↔ multi-year labs ↔ sleep patterns to surface what no single source could see.
Built as a medical aid, not a replacement
- Investigation exports — overlay every signal in one packet a clinician can scan in minutes, supporting their diagnostic decisions.
- Questions to ask your doctor — generated with rationale and source data, so visits become targeted and personalized.
Why we win
Privacy is architectural, not aspirational.
01 · Storage
Two-tier encrypted vault, on the user's machine.
Tier 1 holds full PHI, encrypted column-by-column with AES-256-GCM under a KEK/DEK split derived from the user's passphrase. Tier 2 is a one-way pseudonymized mirror — the only surface visible to cloud AI.
- Cipher
- AES-256-GCM
- KDF
- Argon2id
- Password rotate
- re-wrap DEK in ms
02 · Egress
Deterministic PHI firewall on every outbound call.
A regex-based firewall with measured 99.92% recall, hardened by clinical BERT and RoBERTa NER models as advisory cross-checks. ML never makes the privacy decision — the gate is deterministic. Any layer detecting PHI hard-blocks.
- Primary
- regex (deterministic)
- Recall
- 99.92%
- Fallback
- BERT · RoBERTa · heuristic
03 · DNA privacy
Your DNA is your fingerprint and stays on your device.
Genomes can never be revoked or changed. Further parses your raw file locally and reads only the SNPs needed for analysis. Your raw data is never uploaded, never logged, never shared. Same SNP, same finding, every time.
- Parsing
- local-only · raw file never leaves device
- SNPs
- 711 read · 81 active risks
- Pharmacogenomics
- 229+ drugs across 19 enzymes
“The deterministic core is already shipped. KEK/DEK migration, PHI firewall, and the HIF coordinator are all live — engineered, not described.”
Competitive landscape
Every incumbent owns one or two slices. No one owns the synthesis.
Capability | WHOOP Advanced Labs Wearables + labs | InsideTracker · Function Lab-centric | 23andMe DNA-centric | Apple Health Distribution | Further Health Synthesis |
|---|---|---|---|---|---|
| Wearables + labs + records + DNA | ~ | ~ | |||
| Cross-signal hypotheses with confidence | ~ | ||||
| Local-first, end-to-end encrypted vault | ~ | ||||
| Deterministic genetic + Rx engine | ~ | ||||
| Action layer · labs · prescriptions | ~ | ||||
| No third-party data brokers | ~ | ~ |
Market opportunity
Three tailwinds converging on the unified layer.
- 600M
Wearable units sold annually
and growing double-digit YoY across Apple, Oura, WHOOP, Garmin.
- $18B
At-home diagnostics market by 2030
Quest, LabCorp, and direct-to-consumer panels expanding into preventative care.
- 79%
of consumers concerned about how companies use their data
privacy is now a buying criterion, not a soft preference.
Wearables & at-home diagnostics are now mainstream.
The supply side of consumer health data has crossed the chasm. The synthesis layer hasn't.
Consumers want preventative, personalized care.
The market for paying out-of-pocket for clarity (functional medicine, longevity clinics) is already proven.
Privacy is regulated and expected.
FTC enforcement against BetterHelp, Flo, and GoodRx has set a new bar. Architectural privacy is no longer optional.
Business model
Seven expansion paths from the same engine.
Phase 01
Paid Deep Dive
One-time personal health analysis. High-margin, no infra dependency.
$19.99
Phase 02
Subscription
Ongoing insights, trend tracking, recurring HIF analysis.
$25–$50 / mo
Phase 03
Premium tier
Concierge service: a dedicated virtual doctor with a set number of visits per year. AI synthesizes every signal so each check-in is fast, evidence-rich, and ready for lab or medication orders.
$100–$300 / mo
Phase 04
Diagnostic referrals
Quest, LabCorp, Nucleus, etc. The highest-intent leads in the market.
Affiliate margin per order
Phase 05
Clinical layer
HIMS-style telehealth scripts backed by real longitudinal data.
Per-script + recurring
Phase 06
Clinician API
Doctors get higher-quality, longitudinal patient profiles than today's systems can show — with patient consent. Opens the institutional wedge.
Per-seat + per-pull
Phase 07
B2B engine
License the intelligence engine + PHI firewall to clinics and longevity programs.
Enterprise contracts
$19.99
Day-1 SKU
We provide paid analysis on day one. No infrastructure dependency, no scale required to charge. Every later phase compounds on the same vault, the same firewall, the same engine.
Get on the day-1 listTeam
A neurobiologist-operator and an AI engineer who ship.
Co-founder & CEO
Zade Kal
UC Davis · Neurobiology, Physiology & Behavior · EE + Tech Mgmt minors
Business-development and product operator across fintech, digital assets, and healthcare. Led GTM, partnerships, and integrations for a consumer fintech card at Sophon Labs.
- Project Manager at Penumbra — enterprise system implementations (Salesforce), process mining (SAP Signavio), GMP operations.
- Clinical Supply Engineer at Genentech — GMP production for Phase 1–3 trials; managed SAP master data across 45+ drugs and built a 2–10 year capacity forecasting tool.
- Shipped Siftly (open-source local-first knowledge tool), a tutor marketplace, and several workflow-automation products.
Sophon LabsPenumbraGenentechUC Davis
Co-founder & CTO
Sunney Kang
Washington University in St. Louis · BA Mathematics
AI Engineer at McCarthy Building Companies. Built and scaled a RAG-based Change Order Agent driving 7× the submission volume of top human performers — ~$138K in measured savings.
- Designed a Palantir Foundry + React/TS app using AI for SaaS legal, risk, and IT-security review. Advises leadership on enterprise AI architecture and data governance.
- Prior data engineer: cut $98K in annual Azure compute, dropped data latency 33%, trained an XGBoost labor-demand model at 97.1% accuracy.
McCarthyImpossible SensingWUSTL
We're looking for accelerator partners who understand trust, deterministic AI, and category-defining wedges in healthcare.
Reach the team
zade@further.healthGet in touch.
We're talking to investors, accelerator partners, design partners, and people who'd find this useful. If that's you, say hi.