fastlane/metadata/android/en-US/full_description.txt (1267 bytes)
1 Subrep makes live captions of the sound of your device. Whisper makes them on the phone, and a share link gives them to other people. 2 3 <b>How it works</b> 4 <ul> 5 <li>The app captures the sound of the apps on the phone (a video, a game, a browser; Android 10 or later) or the microphone.</li> 6 <li>Whisper, on the phone, turns the sound into captions. No sound leaves the phone.</li> 7 <li>The captions go to your share link on honjimaku.com, where viewers read along, and to SubRead Overlay over the player, where a tap on a word opens the dictionary.</li> 8 </ul> 9 10 <b>Speech models</b> 11 The app downloads the model that you choose from Hugging Face (ggerganov/whisper.cpp): tiny, base (82 MB) or small (264 MB). Base keeps up on most phones. 12 13 <b>Cloud captions (optional)</b> 14 Cloud captions use Whisper large-v3 turbo on DeepInfra, through subread.space. They cost hours that you buy on subread.space. The app sends only speech, and only when you choose the cloud. 15 16 <b>Permissions</b> 17 <ul> 18 <li>Screen capture and the microphone permission: for the sound of other apps. Android asks each time the captions start.</li> 19 <li>Internet: for the model download, the share link and the cloud captions.</li> 20 <li>Notifications: for the notification with the "Stop" button.</li> 21 </ul>