Open-source AI for ultrasound diagnostics.
WELES Open is a non-profit project building transparent, clinically-validated AI tools for ultrasound — free of corporate licenses, open to every hospital and clinic.
We bring together clinicians and engineers to build ultrasound AI that anyone can inspect, verify, and improve — every algorithm, every decision, fully open on GitHub.
No corporate roadmap decides what gets built next — the people using the tools and the people building them do, together, in the open.
How WELES Open works
1. Upload & first read
A clinician uploads an ultrasound image and records the first clinical assessment — pathology or its explicit absence.
2. Independent second review
Every image, regardless of the first read, goes to a second, independent physician — quality control that catches misses on both sides.
3. AI-assisted detection
Verified images train and validate our YOLO-based detection models for organ- and pathology-specific analysis.
Build tools you'd actually use.
Tired of expensive, opaque diagnostic tools? Help us train algorithms on real, clinically verified cases and shape how AI supports ultrasound — instead of just receiving it from a vendor.
-
Patient safety first
Every model is trained and validated by practicing physicians.
-
Shape future standards
Have real influence, not just adopt what a vendor ships.
-
Grow academically
Network with the medtech community and publish open-access research.
Write code that saves lives.
Real-time ultrasound image and signal processing is a serious engineering challenge. We're building from scratch — no legacy code, no corporate debt, just clean architecture and open review.
-
Unique, hard data
Real-time medical imaging is the ultimate ML/CV challenge.
-
Zero corporate debt
Pure open source, built from scratch, reviewed in the open.
-
A portfolio that opens doors
An open-source medical AI project is a serious resume line.