Loading...
Today we read your lab reports. Tomorrow we will help you see what nobody else can: hidden risks, silent trends, and the tests that could change everything.
Actively building
Our AI models are being refined daily — so that every value on every report is captured perfectly, regardless of format or quality.
We are training next-generation AI models specifically on the hardest cases: complex multi-column tables, handwritten doctor annotations, and the dozens of unique formats used by Spanish laboratories. Our goal is simple — zero missed values, zero manual corrections.
More detail in every response — reference ranges, severity flags, and seamless connections between clinical data points.
We are expanding the structured output to include reference ranges alongside every value, automatic severity indicators for abnormal results, and tighter connections between all data in the report. The goal: your systems receive everything they need in a single response, with nothing left to interpret manually.
Advanced image processing that turns blurry, tilted, or poorly-lit photos into perfectly readable documents.
Not every report arrives as a clean PDF. Some are photographed with a phone, folded, or faded. We are building intelligent image correction — automatic straightening, contrast enhancement, and noise removal — so that even the worst-quality document produces clinical-grade results.
Coming this year
From a single blood test, identify the probability of developing up to 500 diseases — years before the first symptom.
Imagine detecting that a patient is heading toward diabetes three years before their blood sugar crosses the threshold. Or identifying early signals of kidney disease while there is still time to intervene. From as few as 10-15 biomarkers in a routine blood test, our AI aims to calculate the probability of developing up to 500 conditions — from cardiovascular risk to early cancer signals. Similar models have already been validated in large European hospital cohorts with clinically significant results. Each prediction includes a clear explanation of which values contribute to the risk and what steps to consider next. Designed to give clinicians a head start that could change outcomes.
With as few as 3 blood tests from different dates, AI reveals trends invisible to the naked eye — and predicts when critical thresholds will be crossed.
A single blood test is a snapshot. But three or more tests over time tell a story that no human eye can read unaided. Our AI is designed to detect the slow, silent patterns: kidney function declining by 2% each year, hemoglobin dropping so gradually that no single test triggers an alarm, blood sugar accelerating steadily toward the danger zone. It aims to tell you exactly when a value will cross the critical threshold if no action is taken — giving you and your patient the time to change the trajectory.
AI identifies the follow-up tests that could uncover something important that has not been investigated yet.
A doctor cannot order every possible test. But certain results strongly suggest that something else should be checked. If TSH is elevated, the next step should be Free T4 and Anti-TPO antibodies. If hemoglobin is dropping, a complete iron panel may reveal the cause. Our AI is designed to cross-reference results against thousands of clinical studies and suggest the specific additional tests that could uncover a hidden condition — the tests your patient needs but that nobody thought to request.
On our horizon
Full alignment with the European Health Data Space (EHDS) regulation — well ahead of the 2029 deadline.
Europe is building a unified health data ecosystem. We are preparing MedExtract to meet every requirement: standardized data exchange formats, cross-border interoperability, and robust data governance. When the EHDS becomes mandatory in 2029, our clients will already be compliant.
Proactively meeting the requirements for high-risk AI systems under European regulation.
The EU AI Act sets new standards for medical AI. Rather than wait, we are already evaluating and implementing the requirements: risk management frameworks, data governance controls, full technical documentation, and human oversight mechanisms. Our goal is to be compliant before the regulation demands it.
Pursuing Class IIa medical device certification under the EU Medical Device Regulation.
We are exploring the pathway to CE marking under EU MDR 2017/745 as a Class IIa medical device. This certification would enable MedExtract to be used within regulated clinical decision support workflows — a significant milestone for hospitals that require certified tools in their care pathways.
Expanding to understand lab reports from Portugal, France, Germany, Italy, and beyond.
Every country writes lab reports differently — different formats, units, terminology, and reference ranges. We are building support for European lab reports country by country, starting with Portugal, France, Germany, and Italy. The vision: a single platform that reads any lab report in Europe, regardless of language or format.
Discover health trends across your entire patient population — fully anonymized and GDPR-compliant.
What if you could see the health trajectory of an entire community? Anonymized, aggregate analytics designed to reveal population-level patterns: rising diabetes prevalence in a region, declining iron levels in a demographic, emerging trends that deserve public health attention. All fully anonymized and aggregated in strict GDPR compliance.
Native connectivity with the electronic health record systems you already use.
We are building direct integrations with the major clinical systems used across Spanish and European hospitals. The goal: zero-configuration data flow between MedExtract and your existing infrastructure, so structured lab data appears in your system automatically — no manual import, no middleware, no delays.
These capabilities are live today — trusted by clinics and hospitals across Spain.
From pristine digital PDFs to crumpled photos taken on a phone — our AI reads them all. No manual data entry, no errors, no delays. Just accurate results in seconds.
Every result arrives in FHIR R4 format with standardized LOINC codes and normalized units — ready to flow directly into your clinical systems without a single line of custom code.
Castellano, catalan, euskera, gallego — we understand them all natively. Regional terminology, local number formats, all 17 autonomous communities. Your reports, your language.
No database. No disk writes. No stored patient data. Every report is processed entirely in memory and returned in the response. Full GDPR compliance by design, not by policy.
Deploy on your own servers with full Docker and Kubernetes support. Auto-scaling, monitoring, and health checks included. Patient data never leaves your network.
The most comprehensive medical test dictionary for Spanish healthcare. Over 27,000 entries ensure that virtually every test name — however it is written — is correctly identified and standardized.
One API call. Less than one second. Complete structured results with standardized codes, reference ranges, and clinical-grade output. Fast enough for real-time clinical workflows.
Over 280 built-in physiological range checks catch impossible values before they reach a clinician. If a result does not make medical sense, it gets flagged automatically.
Start with our free tier today and be among the first clinics to receive new capabilities as they launch. Or talk to our team about priority enterprise access.