Analyze My Knee
Conditions Guide
Upload your MRI or X-ray as a ZIP file and our AI consortium of 4 models will analyze it for these conditions. All processing happens in your browser — your files never leave your device.
Upload your MRI or X-ray DICOM files for private, AI-powered analysis.
Start AnalysisOur AI consortium analyzes Knee MRI for 13 conditions including ACL tears, meniscus injuries, osteoarthritis, cartilage damage, and fractures. The four-model consortium cross-references multiple imaging planes to flag findings. Results are informational only and not a clinical diagnosis.
Accuracy depends on image quality, scan type, and condition. Our multi-model consortium (4 independent AI models synthesized by Claude Opus) reduces single-model errors through cross-validation. For clinical decisions, always consult an orthopedic specialist or radiologist.
Yes. No account or registration is required. You select DICOM files from your device, they are parsed entirely in your browser, and rendered images are sent to AI models for analysis. Your raw files never leave your device.
No. The analysis is an educational aid to help you understand your imaging findings before or between specialist appointments. For diagnosis, treatment planning, and any medical decisions, consult a qualified orthopedic surgeon or radiologist.
Proton density fat-saturated (PD-FS) sequences on 1.5T or 3T scanners produce the most reliable results. Standard clinical protocols with sagittal, coronal, and axial planes give the AI consortium the most information. Lower-field or motion-degraded scans may reduce detection confidence.
The AI models were primarily trained on adult imaging. Pediatric studies involve open growth plates and different anatomy that may not be reliably interpreted. Pediatric Knee imaging should always be reviewed by a specialist experienced in pediatric musculoskeletal imaging.
Post-surgical imaging can be analyzed, but metallic implants, hardware artifacts, and altered anatomy reduce reliability. The AI will flag what it detects but confidence scores should be interpreted cautiously. Always share post-surgical scans with your treating surgeon.
Your raw DICOM files, ZIP archives, and original pixel data are processed entirely in your browser and never uploaded to our servers. Only rendered JPEG frames needed for AI analysis are transmitted over an encrypted connection, and they are not stored after the analysis is complete.