ehidome, the clinical NLP toolkit that keeps its distance from the cloud

Field notes

The homepage presents a pipeline for converting clinical free text into entities, de-identified records, and FHIR R4 resources, with examples covering medications, diagnoses, dosage, and frequency. It names both terminology targets—including ICD-10, SNOMED CT, LOINC, RxNorm, UMLS, and CMS-HCC—and deployment paths ranging from Python CPU/CUDA to Apple MLX, Android ONNX Runtime, WebGPU, and REST. The claims are presented as product capabilities; the page does not show implementation documentation or licensing details.