About
BloodGPT.

We build AI infrastructure for clinical labs, clinicians, and healthcare integrators. Our platform turns fragmented lab data into clinician-ready, FHIR-compliant outputs.

Book a demo →Become a partner

3Co-founders

5Live in 5 countries

2024Founded

Our mission

Standardize the most quantitative signal in medicine.

Lab data is the most quantitative signal in medicine, and the least standardized. Every lab encodes results differently. Every LIS speaks its own dialect. Every clinician spends time decoding the same biomarkers from a dozen formats. The result is friction at every layer of the healthcare stack.

BloodGPT exists to remove that friction. We normalize lab data against 41,000+ LOINC codes and 30+ naming variations per biomarker. We output clinically interpretable results in FHIR R4. We stay model-agnostic, so labs, clinics, and integrators choose the AI layer that fits their stack, not ours.

Our customers are labs digitizing their reporting, clinicians who need fast and reliable interpretation, and partners building healthcare products that depend on structured lab data.

41,000+LOINC codes normalized

30+Naming variations per biomarker

FHIR R4Native output format

We built BloodGPT to turn one of healthcare's most underused assets — laboratory data — into clear, actionable insight. Our platform makes lab results more understandable for patients, more useful for clinicians, and more powerful for healthcare systems.

Jonathan KronCEO & Co-founder, BloodGPT

Founders

Meet the co-founders.

Engineers, operators, and product leaders who have spent careers at the intersection of healthcare data and AI.

Jonathan Kron

CEO, Co-founder

London, United Kingdom

Vasilii Lazuka

CTO, Co-founder

Limassol, Cyprus

Nata Savaścienka

CPO, Co-founder

Warsaw, Poland

Principles

How we work.

Four principles that shape every decision we make — from which integrations we ship to how we handle a single ambiguous biomarker.

Clinical relevance first

Every product decision is reviewed against clinical relevance before it ships. If a feature does not help a clinician make a better call, it does not ship.

Deterministic where it counts

Value extraction and threshold logic run on a deterministic pipeline. Same input, same output. We use LLMs where they help — never where reproducibility matters.

Model-agnostic by design

Labs, clinics, and integrators choose the AI layer that fits their stack, not ours. OpenAI, Claude, on-prem — we meet you where you are.

Privacy is the product

No sharing. Your data is never used to train models. Compliance is the floor, not the ceiling — HIPAA, GDPR, SOC 2, HL7.

The team

A team built around medical expertise.

BloodGPT is built by engineers, clinicians, and operators. Our team includes an in-house medical advisor, engineers with medical education, and advisors with NHS, pharma, and healthcare informatics backgrounds.

In-house medical advisorNHS advisorsPharma backgroundHealthcare informaticsFHIR / HL7 specialistsML engineers

Meet the full team →