maziyarpanahi / openmed

Local-first healthcare AI platform delivering 1,000+ clinical NER and PII de-identification models that run entirely on-device for HIPAA-compliant medical text processing.

活跃维护 Apache 2.0 Python Tracked
5.2k 655 3 天前
CIDockerPyPI
healthcare llm on-device on-premise sovereign-ai mlx swift ios swiftui ner pii pii-detection clinical-nlp hipaa local-llm nlp python skills android javascript

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

OpenMed is a local-first healthcare AI platform providing clinical Named Entity Recognition (NER) and HIPAA-compliant PII de-identification that runs entirely on-device. It offers 1,000+ specialized medical models across 15 languages, supporting Python, Apple MLX, Swift/iOS, and browser-based deployment via Transformers.js.

为什么值得关注

OpenMed addresses a critical gap in healthcare AI by enabling clinical NLP and PII de-identification without sending patient data to the cloud, directly tackling HIPAA compliance and data sovereignty concerns. Its breadth of 1,000+ models, multi-platform support (Python, Swift, MLX, WebGPU), and strong community traction with 4,265 stars make it a significant project for privacy-preserving healthcare AI.

优势

  • Comprehensive model library with 1,000+ medical models across 15 languages and 247 PII checkpoints
  • True local-first architecture ensuring HIPAA compliance and zero patient data leaving the network
  • Multi-platform support including Python, Apple MLX, Swift/iOS, and browser-based Transformers.js/WebGPU deployment
  • Strong engineering practices with CI, tests, Docker, 22 releases in 6 months, and 48 contributors

局限性

  • 381 open issues suggest potential stability or feature completeness concerns despite active development
  • Healthcare NLP accuracy and model quality may vary across the 1,000+ models without clear benchmarking transparency

使用场景

  • Clinical named entity extraction from electronic health records (EHR) on-premise
  • HIPAA-compliant PII de-identification of clinical notes and medical documents
  • Building privacy-preserving healthcare AI applications for iOS/macOS using Swift and MLX
  • Browser-based clinical text processing without server infrastructure
目标用户: Healthcare AI developers, clinical NLP researchers, health IT teams building HIPAA-compliant applications, and organizations requiring on-premise medical text processing
学习曲线:
分析模型:LongCat-2.0 | 分析时间:1 个月前