How to pull highlights from one video.

Get the recording

Use one file with everyone in it: a Zoom or StreamYard recording, a Riverside mixdown, or a camera with a room mic. Video or audio-only both work. Speaker Cut reads it where it is and never uploads it.

Set up your show profile

A show profile tells Speaker Cut who hosts the show and what it's about. Add the show's name, your hosts (full names, separated by commas, spelled the way you want them in the transcript), plus its tone, audience and topics.

You can create one right from the upload screen with ＋ New show…, or any time from Manage shows.

Choose Highlight Cut

On the Speaker Cut home screen, pick Highlight Cut. Give the session a name: the date and the guest works best, like 2026-09-14 · Sam Rivera. Speaker Cut uses the guest's name from the session name to spell it correctly.

Pick the show and the head-count

Choose your show in Show profile. People in it is then set to your hosts plus one guest. Change it if this episode is different. Auto works for most recordings, but if two people have similar voices, picking the exact count keeps them apart.

Choose the recording and find who's talking

Click Click to choose the recording, pick the file, then Find who's talking. Speaker Cut listens to the whole recording, separates the voices, names each one from the conversation, and writes a word-by-word transcript.

Listen to each voice and fix any name

The Voices card lists everyone Speaker Cut heard, with the name it chose, the person's role, how sure it is, and how long they talked. Play the ▶ samples beside each voice. If a name is wrong, type the right one and click Save names. The transcript and AI notes use these names.

Point out whose voice a clip is

If a voice mixes two people, play its samples and pick whose voice each one is, or move a clip to the right person. Then click Re-separate using your picks. A line only moves when it sounds clearly closer to one of your picks, so the rest of the transcript stays put.

Fix lines right in the transcript

The transcript edits like a document, and every change is saved as you go:

• Change who's speaking: click the speaker's name on a line and pick the right person. Speaker Cut offers to fix the other lines that sound like the same person. Click Change or Keep as is.
• Split a line: click the word where the new speaker starts and press Enter, then choose who said the rest.
• Listen: click any word and press Space to hear a few seconds from there, with the video preview following along. ← → step word by word.

Generate the notes and highlights

Open Notes & Highlights. In What are you looking for?, describe what you're after, or leave it blank for the strongest moments overall. Then click Generate summary, notes & highlights.

You get a summary, timestamped show notes, highlights and cold-open hooks. Each highlight tells a complete story, with a setup, a turn and a payoff, stitched from anywhere in the episode. Every piece starts and ends on a whole sentence.

Read why each moment made it

Why these picks sums up what the AI looked for and which strong moments it left out. Click a highlight's title to open it: you'll see its real length across all of its pieces, why these lines were chosen, and the transcript of every piece in play order.

Refine any clip with a note

In an open highlight, write what you want in Tell the AI what to change and click Submit & re-roll. Afterwards, What changed shows how the AI read your note (expand, tighten, rework or replace) and what it did.

• Expand on this keeps your lines and builds around them.
• Tighten it to under a minute trims but keeps the story whole.
• Add three lines from Sam adds lines Sam actually says, and Speaker Cut checks it really is him talking. If it comes up short, it tells you.

Build the Final Cut package

Below the hooks, set Cut silences over and, if you like, Remove filler words. Then click Build Final Cut package → one FCPXML and import that file into Final Cut Pro. You get one event with a Highlights reel, a Cold-open reel and a project for each clip, all linked to your original recording.

Need the words elsewhere? The transcript downloads as .txt or .srt captions, and Export PDF saves an episode report with the summary, show notes, highlights and hooks.