I built this because iOS will happily show you captions while you watch and then drop them the moment you share the file. The first version shipped in August. Since then the thing that changed most is the editor: 26 templates, karaoke highlight in four styles, keyword emphasis, a filler-word remover, and word timings that survive when you fix a word (an LCS realigner, so editing "teh" to "the" no longer resets the whole line). Smart placement looks at each segment for faces and on-screen text and moves captions out of the way. iPad arrived in 2.3, with a landscape split view. Everything runs on the device. There is no server in this app: transcription uses Apple's SpeechAnalyzer, translation uses the system framework, and the new AI Assistant in 2.7 (keyword picking, title ideas, emoji) uses Apple's on-device Foundation Models. That last one needs an iPhone with Apple Intelligence, and the menu says so instead of failing quietly. The honest limits: accuracy is Apple's model, excellent in the languages it covers and basic elsewhere, and there are no speaker labels because the framework does not provide them. It is free to try, then a one-time purchase. Happy to answer anything about the on-device speech and language APIs; there still is not much written about them.

CapScribe

Captions burned into your video, entirely on your iPhone

Captions burned into your video, entirely on your iPhone