3,000 Word

The problem

Vocabulary lists alone provide limited support for understanding usage, practising pronunciation, and remembering words over time. Thai learners need a manageable daily routine with Thai-language guidance and a way to revisit difficult words.

What Evolext built

• A Thai-language learning web application with 3,000 vocabulary entries and 30,000 example sentences.
• Thai meanings, simple English definitions, IPA pronunciation, and Thai reading guides.
• Word and sentence audio through available device voices, with adjustable playback speed.
• Daily practice with revealable answers and spaced review based on self-reported recall.
• A searchable vocabulary library with difficulty, part-of-speech, bookmark, and learning-status filters.
• Progress tracking, resumable lessons, local browser storage, and JSON backup and restore.
• Installable web-app configuration and service-worker caching for offline access after initial loading.

Engineering approach

The application uses React, TypeScript, Vite, IndexedDB through Dexie, browser speech synthesis, and service workers. Learning progress is stored in the current browser, while JSON export and restore provide a user-controlled way to move or recover that local data.

Practice and review flows use self-reported recall to schedule later review. This creates a repeatable study workflow, rather than a measurement of whether vocabulary has been retained.

Observable output

A vocabulary-learning web application containing bilingual study materials, interactive practice sessions, and locally saved learning progress. No measured improvement in vocabulary retention or learning outcomes is claimed.

Evidence and limitations

The implementation and bundled content support the listed capabilities. All 3,000 entries are currently marked as drafts, with no entries recorded as human-reviewed; meanings, examples, and pronunciation guides require bilingual editorial verification. The documented Oxford source contains 2,998 entries, so the application’s 3,000-entry library should not be described as an exact reproduction. Audio availability depends on device voices, and progress remains in the current browser unless exported. Public adoption and educational effectiveness are not established.