DOI: 10.11118/ecos2026_016

CZECHUP: A WEB SYSTEM FOR PREPARING FOREIGNERS FOR THE CZECH LANGUAGE EXAM WITH ADAPTIVE VOCABULARY AND AI-GENERATED EXERCISES

Polina Komárková¹

¹ Department of Economics, Faculty of Business and Economics, Faculty of Business and Economics, Mendel University in Brno, Zemědělská 1, 613 00 Brno, Czech Republic

This article deals with the design and implementation of the CzechUP web system that helps foreigners prepare for the mandatory Czech language exam required for permanent residence (level A2) or citizenship (level B1). The work responds to the lack of suitable tools for Russian- and Ukrainian-speaking users, for whom the most popular application Duolingo is available only via English. The system integrates a database of 4,372 words parsed from three official CEFR textbooks, the Majka morphological analyzer (up to 347 forms for a single adjective), exercise generation via the Gemini API, and translation via Seznam Slovník and DeepL. The backend is built in ASP.NET Core over PostgreSQL, the frontend in Next.js with localization into three languages. User testing with five respondents confirmed the usefulness of the grammatical-form filter as a unique feature and estimated the error rate of AI-generated exercises at approximately 10%.

Klíčová slova: Czech language teaching, Czech for foreigners, CEFR, web application, morphological analysis, Gemini API

stránky: 17-17, online: 2026