// clinical_nlp_extractor
Paste clinical text. Get structured entities — diagnoses, medications, dosages, lab values, vital signs, and dates — extracted and highlighted in real time.
// clinical_text_input
Presets:
// entity_legend
MEDICATION
DIAGNOSIS
DOSAGE
DATE
LAB VALUE
VITAL SIGN
// annotated_output
// annotated text will appear here after extraction
$ cat README.md
This demo uses a rule-based NLP pipeline: regex patterns for dates, dosages,
vital signs and lab values; curated medical vocabularies for medications (250+
drugs) and diagnoses (300+ conditions). No ML model is called at runtime —
extraction is deterministic and runs in <50 ms per note.
$ cat production_note.md
The production system replaces the vocabulary matcher with a fine-tuned
BioBERT model and a spaCy post-processor, trained on de-identified EHR data.
Output is validated against SNOMED CT and mapped to FHIR R4 resources.