The speech model now comes with the app. Nothing to download, nothing to set up: grant the microphone and say your first entry, in Ukrainian, English or the mix of them, on any connection or none. Also in this version: • Onboarding is three short screens. • The Speech section is gone from your profile: there is nothing left to choose there. • The app is larger to install because the model rides inside it; updates stay small. • If the connection drops while you are naming an entry, the entry is saved with your word and Vidlun's reading arrives once you are back online. • The evening reminder no longer arrives twice.

Speak English? Vidlun now understands it with your iPhone's own on-device recognition, so there is nothing to download before your first entry. Ukrainian and mixed speech keep Vidlun's own model, which you download once - now from a new Speech section in your profile, where you can also change the language you speak or delete the model. Also in this version: • Shorter onboarding and clearer permission screens. • The model download is more resilient: if Apple's servers can't deliver it, Vidlun fetches it directly. • A typed entry now says it has no audio instead of claiming that recordings are switched off. • Layout respects the iPhone's side safe areas.