| Information / permission | When | Purpose | Necessary? |
|---|
| Camera video | Talk-through / single-cam scenes / photos | Compose the video or save a still | Necessary when you use the feature |
| Microphone audio | Recording | Talk-through audio; cartoon lip-sync; teleprompter pace-follow | Necessary when you use the feature |
| Speech recognition / on-device models | Captions | Transcribe the clip on device into editable captions; audio is not uploaded to our servers. iOS may download an Apple model once if a pack is missing; Android may download an on-device model the first time you use a language (model files only) | Necessary when you use the feature |
| Caption text (on-device cache) | After adding captions | Let you edit again and export a captioned video; Clear cache does not delete this | Necessary when you use the feature |
| Photo library read/write | Save video and photos / import scene media | Write and open videos and photos | Necessary when you use the feature |
| While-in-use location / place name | Info watermark place is on | Reverse-geocode on device and write the watermark; coordinates are not uploaded to our servers; the place cannot be edited | Necessary when place watermark is on |
| Coordinates (about 4 decimal places) / temperature and conditions | Weather card is selected | Sent to Open-Meteo for current weather on the watermark; does not go through our servers; weather cannot be edited | Necessary when you choose the Weather card |
| Teleprompter script | Teleprompter | Store cue text on device (up to 10 scripts); not in the video, not uploaded | Necessary when you use the feature |
| Topic, notes, or draft script | AI script, polish, or expand | Sent through our servers in mainland China to Volcengine Ark (Doubao) to generate or rewrite text; the result returns to the device. We do not keep the script body and do not use it to train models | Necessary for AI script |
| Daily AI-script count | After you use AI script | Enforce the 40-per-day quota | Necessary for this feature |
| Narration script | Generate voiceover with a cloud voice | Sent through our servers to MiniMax to synthesize speech; audio is returned to the device. We do not keep the script text. The default on-device voice does not upload the script | Necessary for cloud voiceover |
| Your own voice sample | Record “My voice” | Sent through our servers to MiniMax to create your narration voice. Only a voice you confirm is yours; we do not keep the audio file long term | Necessary for voice cloning |
| Personal-voice ID, name, created and deleted times | After you buy or record a personal voice | Show, fulfill, and support the voice. After you delete it in the App, the record is kept until you delete the account | Necessary when you use the feature |
| Cloud-narration usage, extra packs, and expiry | Cloud voiceover / extra characters | Fulfill monthly quota and extra packs; unused extra credits expire | Necessary when you use the feature |