Under the hood

What CutPrep actually does between hitting Run and the timeline opening.

Six steps, plain language, almost no jargon (FCPXML slips in once or twice). Skim if you're not the curious type. If you ever sit next to an editor who asks "how does this thing actually work?", this is the page to send them.

①

Read what was said

CutPrep starts by reading the transcript you handed it (an SRT or VTT exported from Whisper, MacWhisper, or your own service). The transcript is the map: every quote, every laugh, every hesitation lives in there with its exact in- and out-timecode. We never touch the original video file at this stage.

00:14:22 → 00:14:31
"The interesting part wasn't the launch itself. It was the year we'd spent learning what we actually wanted to build."
00:18:05 → 00:18:14
"When you've been at it for a decade you stop chasing every brief. You pick the ones where you can do your best work."

If your transcript is clean, everything downstream is clean. If it's messy, well. Garbage in, garbage out, just like the edit.

②

Read what you asked for

Next CutPrep reads the brief: which preset (podcast, documentary, branded content, …), which client, which one-sentence intent you typed. That tells the editorial side which formats to build, how long each should be, and which tone the cut needs to sit in.

📁 Footage
├─ cam_A_01.mov
├─ cam_A_02.mov
├─ cam_B_01.mov
├─ audio_01.wav
└─ transcript.srt

The brief is the steering wheel. No brief, no edit.

③

Pick the moments

CutPrep then makes a structured selection per format: the actual editorial work. For each format it picks the timecode ranges that should be in the cut, in the order they should sit in the timeline. This is the step that used to take a junior an afternoon.

# selectie · podcast preset
Topic A snippet 1 · 00:14:22 → 00:14:31 ✓
Topic A snippet 2 · 00:18:05 → 00:18:14 ✓
Topic B snippet 1 · 00:32:11 → 00:32:20 ✓
Flow cut · 12 fragments selected

It's not magic. It's a very fast, very literate first pass. You'll still tighten every cut by hand.

④

Hand your NLE a timeline

Once the selection exists, CutPrep writes a clean timeline per format, as FCPXML for Final Cut and as XML for Resolve and Premiere, so it imports natively wherever you cut. Open the file and the timeline is laid out exactly as the selection said, with every cut in the right place against the right source clip.

TopicA.fcpxml

TopicB.fcpxml

FlowCut.fcpxml

Spotting.fcpxml

No proprietary format, no plugins, no funny business. Standard FCPXML. Standard import.

⑤

Cut the previews so you can scan

To save you from opening your edit suite just to see what got picked, CutPrep also cuts a small MP4 per format: a fast preview of the selected fragments stitched together. You can scrub through it in QuickTime in seconds and decide which formats need a deeper pass before you import.

TopicA.mp4

TopicB.mp4

FlowCut.mp4

Spotting.mp4

Especially helpful when you have five formats waiting and only two coffees of energy.

⑥

Leave a tidy bundle in your folder

Everything lands in one per-run subfolder inside your project: the FCPXMLs, the MP4 previews, an audio sidecar synced to the master timecode, and a plain-text run log. Nothing leaves your Mac. The footage you started with is untouched.

📁 CutPrep / 2026-05-31_1410
├─ 4 × .fcpxml
├─ 4 × .mp4 preview
├─ audio_synced.wav
└─ run-log.txt

Then you close CutPrep and do what only an editor can do.

What we are not

What CutPrep does not do.

It does not need a transcript from you. It makes one itself, on your Mac, so nothing goes out to be transcribed.
It does not transcode, grade or render the final film. Your NLE does all of that; CutPrep hands it a clean starting point. The render bar is still yours to stare at.
It does not upload your footage. Picture and sound never leave your Mac; only text goes out for the thinking part.
It does not run on Windows or Linux. Native macOS on Apple Silicon. That's the whole spec sheet.

Want to play with the actual cost of running this? The calculator is over here.