Every Opus Clip feature explained (for people who have never used a clipping tool)
Opus Clip takes one long recording and returns a batch of short vertical videos with captions burned onto them. Everything else on the feature list is a variation of that: ClipAnything decides where to cut, the virality score ranks the cuts, AI reframe crops each one to portrait and follows the speaker, captions and templates style the text, b-roll drops in stock footage, and the scheduler posts what you approve. The features are real. The judgement about which clip is worth posting is still yours.
First, what a clipping tool actually is
If you have never used one: you upload an hour of podcast or a webinar recording, and a few minutes later you get back somewhere between five and thirty short videos, each roughly 20 to 90 seconds, cropped to a vertical shape for Reels, TikTok and YouTube Shorts, with the words on screen. You did no editing. That is the whole promise, and Opus Clip is the most popular tool that delivers it.
The jargon, translated
| Term | What it means in plain English |
|---|---|
| Clipping | Cutting one long video into several short standalone videos |
| Reframing | Changing a wide 16:9 video into a tall 9:16 one, cropping in on whoever is talking |
| Burned-in captions | Subtitles baked into the video file so they show without the viewer turning anything on |
| B-roll | Extra footage or images laid over the speaker while they keep talking |
| Virality score | The tool's own 0 to 100 guess at how well a clip might do |
| Credits or processing minutes | The billing unit. Roughly one credit per minute of source video you upload |
| Hook | The first line of the clip, the bit that decides whether anyone watches past second one |
ClipAnything: how it decides where to cut
ClipAnything is the part that reads your recording and picks the moments. On the free plan it works from the spoken words alone, meaning it reads the transcript and looks for segments that sound like a self-contained thought. On paid plans it also considers visual objects, sound and emotion, so you can ask it for things like every moment someone laughs or every time a product appears on screen. It is genuinely useful for finding candidates in an hour of audio you do not want to sit through.
- What it is for: turning a long recording into a shortlist in minutes instead of an afternoon
- What it cannot do: know which of those moments your specific audience cares about
- Where it goes wrong for beginners: cutting a thought in half, so the clip ends before the point lands
Virality score: a sorting tool rather than a prediction
Every clip comes back with a number out of 100. Treat it as a way to sort the batch so you watch the promising ones first, and nothing more than that. The score reads structural signals like whether the clip opens with a question, whether the pacing is quick, whether the topic is clearly stated. It has no idea who follows you, what your audience has already seen, or what is moving on the platform this week. Plenty of creators find low-scored clips outperform high-scored ones. Use it to order your review queue, then trust your own ears.
AI reframe: the vertical crop
Your podcast was probably filmed wide, with two people side by side. Reframe crops that to a tall phone-shaped frame and moves the crop to follow whoever is speaking, which is why a clip can cut between faces without you cutting anything. On a single-speaker recording it is close to reliable. On a two or three-person podcast it is the feature most likely to embarrass you: it can lock onto the wrong person, or crop the top of a head off, or drift mid-sentence. Opus lets you drag the frame manually per clip, which works, and which is also more editing than the marketing implies.
Captions and templates
Captions are automatic, synced word by word, and styled by a template you pick. This matters more than beginners expect, because most short-form is watched on mute first, so the words on screen are doing the selling. The transcription is good on clean audio in English. It is weaker on proper names, industry jargon, accents and the smaller supported languages, which Opus covers to roughly twenty-plus with some of them still flagged as beta. Always read the captions before you post. One misspelled name in giant letters is the kind of thing viewers screenshot.
B-roll
B-roll drops stock clips or AI images over the top of you talking, so the screen changes every few seconds instead of showing one static face. The theory is sound, since visual change holds attention. In practice the matching works off keywords rather than meaning, so a clip about running a business can come back with someone jogging. It is worth a look on every export, and it is worth deleting the ones that miss.
AI curation and the rest of the workspace
The remaining features are workflow rather than editing. AI curation groups and filters the batch so you are not scrolling thirty exports. The social scheduler connects your accounts and posts approved clips to Shorts, TikTok, Reels, LinkedIn, Facebook and X on a calendar. Brand kit stores your fonts, colours and caption style so every clip looks like it came from the same show. Team seats let an editor or VA work in the same workspace. On Opus these sit on the higher tiers rather than the entry plan, so check what your tier includes before you plan a workflow around them.
What no feature on the list does
Read the whole feature list back and you will notice the gap. Nothing on it decides whether a clip is worth your audience's time. Nothing rewrites a flat opening line into one that stops a thumb. Nothing checks that the clip makes sense to someone who has never heard your show. Those three jobs decide whether a clip gets watched, and they are the three the software leaves on your desk along with the weekly reviewing and posting.
So, tool or done-for-you
If you have never made a clip, get the free plan, run one episode through it, and watch what comes back. You will learn more about short-form in that hour than in any guide, and you will see exactly which parts you enjoy and which parts you will quietly stop doing by week three. If the answer is that you will stop doing all of it, hand the whole job to a person instead. At Brandboost you record and we do everything else: pick the moments, write the hooks, caption, reframe, review every clip and post it across platforms. We back it with 20k views in your first week or you don't pay, and we work with 20+ verified clients at a 5.0 rating on Upwork. Book a short growth call and we will map it out for your show.
FAQ
What does Opus Clip actually do?
It takes a long video such as a podcast or webinar, finds short self-contained moments inside it, crops each one to a vertical phone shape that follows the speaker, adds burned-in captions, and gives every clip a score out of 100. You then review the batch and post what you like.
What is ClipAnything?
ClipAnything is the moment-finding engine. It reads your recording and picks segments that look like a complete thought. On paid tiers it can also search visually and by sound or emotion, so you can ask for specific things like laughter or a product appearing on screen.
Is the Opus Clip virality score accurate?
Treat it as a sorting aid rather than a prediction. It reads structural signals like hook phrasing and pacing, and it knows nothing about your audience or the platform this week. Use it to decide which clips to watch first, then judge with your own ears.
Does Opus Clip work for a two-person podcast?
It works, with a caveat. Auto-reframe on multi-speaker footage is the feature most likely to need a manual fix, since it can follow the person who is listening or crop a head. Budget time to check the framing on every clip before posting.
Do I need the paid plan to try it?
No. The free plan gives you roughly 60 processing minutes a month with a watermark on exports and short-lived access to the files, which is enough to run one episode and see whether the workflow suits you. Paid tiers remove the watermark and unlock the visual clipping, reframe controls, b-roll and scheduler.