TL;DR
OpenMed is an on-device AI library that identifies and removes personally identifiable information (PII) from medical text and extracts clinical entities. It supports over 1,000 models and runs locally in Python and Swift environments.
Key features
PII De-identification: Real-time detection and masking of sensitive information such as patient names, addresses, and IDs using 247 PII checkpoints
Clinical Entity Extraction: Extracts structured information such as diagnoses, medications, and test results from text
Multi-platform Support: Same functionality via Python package (pip install openmed) and Swift framework (OpenMedKit) for iOS/iPadOS/macOS
Apple MLX Optimization: Accelerated inference on Apple Silicon, no cloud dependency
12 Language Support: Capable of processing multilingual medical documents
Apache 2.0 License: Free for commercial use and modification
When to use it
When hospitals or research institutions need to de-identify patient data locally without transmitting it externally
When embedding medical text analysis features into iOS apps (e.g., patient note scanning, clinical trial screening)
When developing medical AI applications where on-device processing is essential due to cloud costs or regulations (e.g., HIPAA, GDPR)