Recognition and translation use built-in iOS capabilities, and the first run downloads Apple's English speech assets and the Chinese translation pack. That is the only step that needs the internet — everything works offline afterwards.

Yes. After the one-time setup, recognition, translation, and recording all happen on the device. No connection is needed at all.

It listens to English lectures and translates them into Chinese. Other languages are not supported yet.

Not to us — recognition and translation happen entirely on the device. One thing to note: if iCloud Backup is enabled in iOS Settings, caption text rides along in that encrypted backup to Apple (recordings are excluded). Beyond that, content leaves the phone only when you export it yourself through the system share sheet.

An orange dot means the app hasn't heard anything for a while: move the phone closer to the speaker and keep the room quiet. If it doesn't recover, tap "结束" (Stop) to end the session, then tap the mic button to start over — the captions you already have are safe.