Script and captions
Apple’s speech recognition (the new SpeechAnalyzer on macOS 26). Optional: NVIDIA’s Parakeet model, downloaded once, then run on your Mac.
Finding “um”s
Sound analysis of pitch and tone, using Apple’s Accelerate maths library. Speech recognition alone misses most of them.
Finding restarts
Several checks vote: matching words, words that sound alike, and how your voice rises and falls. Apple’s Natural Language framework helps, and Apple Intelligence can double-check.
Hiding private details
Apple Vision text recognition reads the screen, then pattern checks spot emails, card numbers and keys.
Removing your background
Apple Vision person segmentation.
Smart framing and thumbnails
Apple Vision face detection, plus your recorded pointer and clicks.
Automatic zooms
Your recorded clicks, pointer and typing. No guessing from the picture.
Write my script, Edit by asking, titles, chapters
Apple Intelligence, Apple’s on-device language model (macOS 26). Simpler versions work without it.
Say it in my voice
Apple Personal Voice, which you make and keep on your Mac.
Voiceover
Your Mac’s built-in voices.
Caption translation
Apple’s Translation framework.
Noise removal and voice styles
Sound processing in the app, using Apple’s audio tools.
Recording and saving
Apple ScreenCaptureKit and AVFoundation, with H.264 or HEVC.
Search your library
On-screen words from Vision, spoken words from speech recognition, and Apple’s Natural Language for questions.
