Speaker Cut. Download free
The Highlight Cut guide · One & Pro

One recording in. Whole stories out.

No multicam and no separate mics, just the recording. Speaker Cut tells the voices apart, works out who's who, writes the transcript and finds highlights with a beginning, middle and end, ready for Final Cut Pro.

.mp4 · .mov · .wavSpeaker Cut One & ProFinal Cut Pro 12+Apple silicon · macOS 15.6+
Mic per person?If everyone had their own mic, use the multicam guide instead. With a clean mic per person, Speaker Cut knows exactly who's talking without telling voices apart.
Part 1

Before you start.

Two things make the difference: a recording where everyone can be heard, and a show that knows its hosts.

  1. Get one recording with everyone in it

    A Zoom, StreamYard or Riverside recording, a camera with a decent room mic, a phone on a tripod, or the edited episode. Video or audio-only both work. Speaker Cut reads it where it is and never uploads it.

  2. Set up the show once

    Make a show with its hosts' names, spelled the way you want them in the transcript. You can do it in the first-run setup, from Home, or with New show… right on the Highlight Cut screen. Next time, the hosts are recognized and named for you.

Heads-up: heavy cross-talk and echoey rooms make voices harder to tell apart. Music beds, bumpers and short clips are fine; they show up as stray voices you can ignore.
Speaker Cut — Home
Speaker Cut's Home: four ways in, with Highlight Cut as the fourth card, and your projects grouped by show.

Highlight Cut is the fourth door on Home.

Part 2

Find who's talking.

Point Speaker Cut at the file, tell it the show and the head-count, and let it listen.

  1. Choose Highlight Cut

    On Home pick Highlight Cut, then drop the recording or click Drop one recording, or click to choose.

  2. Name the session

    The date and guest work best, like 2026-09-14 · Sam Rivera. Speaker Cut takes the guest's name from the session name to spell it right.

  3. Pick the show and People in it

    Choosing the show sets People in it to its hosts plus one guest. Auto works for most recordings; if two people sound alike, pick the exact count.

  4. Find who's talking

    Click Find who's talking. Speaker Cut separates the voices, transcribes every word and names who's who, all on your Mac. A long episode takes several minutes; you can do something else while it works.

Speaker Cut — Highlight Cut
The Highlight Cut screen: drop one recording, then Session name, Show and People in it, and Find who's talking.

Session name, Show and People in it sit next to the drop zone.

Part 3

Check the voices.

A quick listen here means every highlight is credited to the right person.

  1. Play a few clips

    Each voice card shows the name Speaker Cut chose, how sure it is, how long they talked and up to three clips to play. Badges tell you where the name came from: Recognized voice, Host, Named from the talk, or A guess. Check it.

  2. Fix any name

    Click a name and type the right one. It saves as you type, and the transcript and highlights use it.

  3. Two people came out as one?

    Under Not quite right?, change the head-count and click Re-separate voices. It reuses the first listen, so it takes seconds, and names you've typed are kept.

  4. Continue to highlights

    Click Continue to highlights.

Illustration with example names.

Part 4

Get your highlights.

Ask for what you want, read why each moment made it, and refine any clip with a plain-English note.

  1. Find highlights

    Type what to look for, if anything (“his advice for first-time founders”, “the funniest moments”), and click Find highlights. You get highlights, cold-open hooks, and chapters from the show notes.

  2. Every highlight is a whole story

    A setup, a turn and a payoff, stitched from anywhere in the episode, starting and ending on whole sentences. Open one to see each piece in play order and why those lines were chosen.

  3. Refine with a note

    ↻ Re-roll, then tell the AI what to change: “tighten it to under a minute”, “add a line from Sam about hiring”. Leave it empty for a different pick.

  4. Find more, your way

    Find another highlight or Find another hook: choose A story, Advice, A hot take, Something funny and more, preview each, and add the ones you like. Each chapter also has + Highlight from this chapter.

  5. Build the reels

    + Add to reel and + Add to cold open, then reorder on Reels. The show learns from the picks you keep.

Speaker Cut — Highlights
The Highlights tab with chapters, a stitched highlight and its pieces, and highlight cards.
Part 5

Fix the transcript like a doc.

  1. Change who said a line

    Click the speaker's name ▾ and choose from Who said this?. Captions, the export and the AI all use the change.

  2. Search and listen

    ⌘F to find a word; click any word to jump the player there and Space to listen.

  3. Turn a passage into a highlight

    Select text and click Make highlight, or Copy with timecodes.

  4. Download

    Captions .srt, Transcript .txt or a Word doc.

Speaker Cut — Transcript
The Transcript tab with speakers, the transcript text and the player.
Part 6

Export to Final Cut.

One import brings every clip in, linked to your original recording.

  1. Pick what goes

    On Export: the full recording, the Highlights reel, the Cold open and a clip per highlight and hook, each its own project.

  2. Send to Final Cut

    Send to Final Cut opens Final Cut and asks which library to add the event to.

  3. Or build the package

    Build package (.zip) adds Notes & Highlights.pdf, Transcript.pdf, captions (.srt), YouTube chapters and a READ ME.

Needs Final Cut Pro 12 or later (FCPXML 1.14). Update Final Cut from the Mac App Store if it says the version isn't supported.
Speaker Cut — Export
The Export tab: what lands in Final Cut, package files and Send to Final Cut.
Good to know

Common questions.

Do I need Speaker Cut One or Pro?

Yes. Highlight Cut uses the transcript and AI, which are part of One and Pro. The free app still builds and cuts multicams and exports them to Final Cut.

What kind of recording works best?

One file where everyone is heard clearly: a Zoom, StreamYard or Riverside recording, or a camera with a decent room mic. Heavy cross-talk and echoey rooms make voices harder to tell apart.

Two of us sound alike and came out as one voice.

Set People in it to the exact number and click Re-separate voices. It reuses the first listen, so it takes seconds, and any names you typed are kept.

Is my video uploaded?

No. Separating voices and transcribing happen on your Mac. Only the transcript text is sent to the AI to name voices and write notes and highlights.

I have a camera and mic for each person. Should I still use Highlight Cut?

Use the multicam workflow. It cuts the video to whoever is talking and gives you the same transcript, highlights and hooks.

Turn your next episode into clips.

Drop in the recording, pick your show, and let Speaker Cut find who's talking and the moments worth sharing.