• Hanzi image practice — practice characters with image mode. • On-demand ML models — TTS/ASR models are no longer bundled; they download when needed.

Upgraded Chinese TTS with a new Kokoro model and fine-tuned 10 male/female voices for clearer, more natural pronunciation. Fixed polyphonic pinyin across classical poetry and lyrics (e.g. 还, 为, 尽, 更, 曲) for more accurate readings and display.

1 3,000+ new Daily Life dialogues — Expanded everyday conversation practice across greetings, shopping, travel, work, and more, with richer A/B dialogue content in every subcategory. 2 Refreshed home & practice UI — Cleaner home screen layout (stats, progress, practice/test modules) and a more consistent look across article and dialogue practice screens. 3 Smoother reading & listening flow — Full-width English translations, relocated “Listen in English” controls (bottom bar in articles; next to A/B speakers in dialogues), tighter spacing, and improved Chinese TTS/pinyin playback.

Improved Chinese TTS and Word Segmentation Enhanced Chinese text segmentation for more natural TTS pronunciation, with better support for practice playback and shadowing. Corrected Poetry Pinyin Accuracy Fixed polyphonic-character pinyin issues in classical poems, improving reading accuracy for poetry and literary content. Refined Practice Experience Updated the UI and interaction flow across word, character, article, and dialogue practice screens for a cleaner and smoother learning experience.