Czech PII Anonymizer & NLP

Wraps Czech academic NLP tools from LINDAT/ÚFAL into six operations: NameTag for 35-language NER with auto-detection, MasKIT plus regex for production anonymization covering 80+ PII patterns (IBAN across 30+ countries, EU VAT for 28 states, US SSN, crypto addresses, API tokens), UDPipe for morphological analysis, PONK readability scoring, Korektor spellcheck, and Charles Translator for eight languages. The anonymization layer includes a zero-egress local mode that keeps all text on your machine. Tested across legal, medical, banking, and real estate documents with international coverage for US, UK, German, French, Italian, Spanish, Polish, Russian, and Indian identifiers. Non-commercial use only due to CC BY-NC-SA model licenses.