Turn raw footage into a rough cut with plain-English instructions.
Passcut uses Apple Foundation Models to convert natural language prompts into precise transcript segment selections without sending video off your Mac.
What is AI rough cut automation on macOS?
Passcut is a native macOS application that turns raw video footage into a frame-accurate rough cut from a plain-English instruction, executing entirely on Apple Silicon without cloud processing or API keys.
By combining Apple SpeechAnalyzer for on-device speech-to-text with Apple Foundation Models for natural language understanding, Passcut translates instructions like “cut this down to the eight minutes about network neutrality and drop the tangents” directly into timeline edits.
How does Passcut translate natural language into video cuts?
Passcut uses on-device language models exclusively to select transcript text segments, while a deterministic audio engine resolves those selections into exact timeline timecodes by snapping to silence and frame boundaries.
Because the language model only evaluates text and never emits timecode numbers directly, cuts cannot drift or skip frames. You review the generated cut segment by segment alongside the assembly timeline before exporting.
What happens after the AI rough cut is generated?
After Passcut generates a rough cut, you can refine the timeline using assembly editing, click transcript lines to seek or reject clips, adjust volume, and export to FCPXML, EDL, or MP4.
Every project is saved locally in a .passcut bundle containing the timeline, transcript, cut decisions, and media references with autosave protection. You can export directly to Final Cut Pro, DaVinci Resolve, or Premiere Pro for finishing.
Automate Your First Cut on macOS
Join the Passcut waitlist to test on-device AI rough cut generation on Apple Silicon.
Requires macOS 26 or later and Apple Silicon. No cloud processing, zero spam.
#Frequently Asked Questions
How does Passcut generate a rough cut from plain English?
Passcut transcribes speech on-device with Apple SpeechAnalyzer, then feeds the transcript to Apple Foundation Models. The model selects relevant transcript segments according to your instruction, and a deterministic engine resolves those into frame-accurate timeline cuts.
Does Passcut send video footage to a cloud AI model?
No. Video never leaves your Mac. Instruction understanding and transcript parsing run locally using Apple Foundation Models on Apple Silicon.
Can I edit the generated rough cut manually?
Yes. Passcut provides an assembly timeline with connected clips, split, trim, reorder, per-clip volume control, B-roll overlay lanes, text lanes, and music lanes with full undo/redo.
Get Early Access to Passcut
Passcut is currently in pre-launch. Join the waitlist to receive early access as soon as the native macOS build is ready.
Requires macOS 26 or later and Apple Silicon. No cloud processing, zero spam.