Named Entity RecognitionService

Rule-based named-entity recognition for English text (patterns and word lists; not a trained model). POST JSON {"text": "..."} to /v1/ner (max 100,000 characters); the response is a JSON object with count, model and entities: each with text, label, start and end. Labels: PERSON, ORG, LOC, DATE, MONEY, PERCENT, EMAIL, URL, PRODUCT. People are found from name patterns, titles and common first names; organisations from suffixes (Inc, Corp, Bank, University...) and known names; dates in ISO, day-month-year and month-day-year forms. Works best on clean, well-capitalised text. Identical input gives identical output on every supplier.