Svezdrav Rešenja DOO is an innovative, research-driven digital health company from Serbia founded as a direct commercial spin-off from pioneering PhD doctoral research in advanced Electrocardiogram (ECG) signal analysis. Built upon years of rigorous academic investigation and continuous practical, real-world deployment, the company has developed ECG4Everybody—a comprehensive, data-driven ecosystem for non-invasive cardiovascular monitoring. By combining proprietary software, edge-AI algorithms, and flexible hardware, we democratize clinical-grade heart care.
Beyond one-way monitoring, our platform has evolved into a multi-sided digital health community integrating crowdsourced waveform validation, expert-led training hubs, and behavioral tokenomics that reward anonymous data donation to open science. Under the academic leadership of founder Dr. Stevan Jokić (see Academic CV and Full Bibliography), our deep commitment to open science and academic rigor is backed by peer-reviewed, high-impact publications validating our methodologies.
Unlike early-stage startups, our solution boasts exceptional real-world validation: 100K+ global downloads across platforms, a curated dataset of 3 Million+ unique biomedical recordings (including large-scale real-world smartphone PPG datasets gathered from over 1 million unconstrained recordings), and a highly active user base generating more than 1,000 daily usages.
We have developed an innovative, ultra-portable ECG device capable of recording up to 6-lead ECG signals. The device features a highly adaptable design, allowing recordings to be taken with or without traditional electrodes, making medical-grade ECG monitoring effortless for everyday users and clinical trials alike.
Our mobile application features a proprietary software module that utilizes a standard smartphone camera to collect high-quality PPG signals. Through advanced smartphone-only camera analysis, our software accurately extracts Pulse and HRV metrics. It evaluates vascular stiffness and arterial health using advanced preprocessing pipelines, including second derivatives (SD), 2-component/3-component Gaussian fitting, and A/B point landmark detection, removing the barrier of expensive hardware.
The backend system handles real-time cloud-based filtering, R-peak detection, and complex time/frequency-domain HRV analysis. It incorporates benchmarked deep learning (CNN, ResNet, MLP) and ensemble machine learning models (XGBoost, Random Forest) for advanced vascular morphology classification and biological vascular age regression.
An open-science module allowing users to securely and anonymously share cardiovascular graphs for community validation, peer-to-peer knowledge sharing, and collaborative waveform analysis.
Interactive hubs where certified wellness and sports professionals deploy personalized programs, validating their real-time impact on the autonomic nervous system via HRV tracking.
A behavioral design rewarding users with tokens for anonymous data donation. Transparency is maintained through state-of-the-art model explainability workflows (such as Integrated Gradients analysis) to map specific PPG morphological features driving classification and regression outputs.
Our underlying algorithms, data processing pipelines, and datasets are rigorously peer-reviewed and open to the international scientific community. In 2026, our comprehensive study detailing the curation of large-scale smartphone photoplethysmography datasets and machine learning benchmarks for vascular assessment was published in the journal Electronics (MDPI).
To support global digital biomarker development and open science, we maintain public repositories containing our preprocessing pipelines, signal normalization code, and benchmark machine learning models.
We offer an unparalleled resource for international consortia looking for a mature digital health platform with a proven track record: