INDEPENDENT RESEARCH
FOR PEOPLE WHO MAKE VIDEO.

FIELD NOTES
VOL. 001 / NOV 2026

CLIPPING AI.
IN THE CUT.

Independent reviews of AI video clipping tools.
Which tools deserve a place in your workflow?
Follow the evidence, from first upload to final export.

Explore the files

The full archive

Browse all reviews

28 research files on AI clip makers, video editors, caption tools and pricing, grouped by what you need to decide.

See the directory
FILE 001 — DOCUMENTATION REVIEWUPDATED 07 OCT 2026

On the record

Vugola’s API is real.
The fine print
is complicated.

Public API. Public MCP server. Credit units explained; publishing coverage still needs checking. We traced the documentation so you know what to verify before a paid trial.

Open the research file ↗

Hands-on tests / October 19 and 29–30, 2026

We tested them ourselves.

Same three Creative Commons sources (a stage talk, an interview and a screen recording) through six tools. Opus Clip and Vugola ran on October 19, 2026, twice each: 12 jobs. Klap, Pictory and Descript ran on October 29. Captions’ first standard-source uploads failed, so an October 30 run used shorter substitutes; a later retry with compressed standard sources is the run we graded. Every graded clip was logged as accepted, repairable or rejected. The frames below are from the October 19 pair.

Opus Clip
73.5 / 100
Vugola
51.6 / 100

Weighted, provisional scores from one tester and one round: clip quality 30, captions 20, automation 20, value 20, usability 10. They are not a stable industry ranking. Vugola won captions in this pair; Opus Clip won the rest. See hands-on results for all six AI clipping tools, including Klap (82.4), Pictory (79.0), Descript (76.8, partial sample) and Captions (clip grades, no overall score).

Stage talk

Vertical Opus Clip export of a stage talk: the speaker framed full height with bold burned-in captions reading “They hear voices or” and the OpusClip trial watermark top left.Vertical Vugola export of the same moment of the stage talk: the speaker framed full height with small white captions reading “of us. They hear”, no watermark.
Opus Clip · 27 clips from this talk: 14 accepted as-is, 13 repairable, 0 rejected. Trial export, watermarked.Vugola · 5 clips from this talk: 0 accepted as-is, all 5 repairable. No watermark.

Interview

Vertical Opus Clip export of the interview: the host, Roma Agrawal, in a yellow shirt in a close crop, bold burned-in captions reading “Pinpoint the cause of” with “cause” in green, and the vertical “Created with” OpusClip trial watermark on the left edge.Vertical Vugola export of the same moment of the interview: the host, Roma Agrawal, in a looser crop with small white captions reading “pinpoint the cause of”, no watermark.
Opus Clip · 24 clips from this interview: 11 accepted as-is, 11 repairable, 2 rejected. Trial export, watermarked.Vugola · 4 clips, all repairable: every one starts mid-sentence. No watermark.

Screen recording

Vertical Opus Clip export of a Chrome DevTools screen recording: the code editor shrunk to a narrow band in the middle of a blurred background, text too small to read.Vertical Vugola export of the same screen recording: the full browser window shown as a thin band across the middle of the frame, code illegible, captions below.
Opus Clip · 26 clips: 2 accepted, both talking-head moments. Code is not readable on a phone.Vugola · 6 clips, all repairable; 354 seconds of output showed the screen as an illegible band.
See the full results ↗

Frames are real exports from our runs, taken at the same source moment for each pair. Per-tool detail: Opus Clip review · Vugola AI review.

Frames from “How can video games boost your mental health?” by Universiteit van Nederland (Dr Max Birk), “Extreme weather with Hannah Cloke” by The Royal Society and “Debugging JavaScript - Chrome DevTools 101” by Google Chrome Developers, all CC BY 3.0; trimmed; clips generated by Vugola and Opus Clip.

From the test bench / October 2026

Real exports, real editors.

Every screenshot below comes from our own runs on Creative Commons sources. Click one to read the page that documents it.

Primer / what we examine

What is clipping AI?

Clipping AI tools turn one long recording into a batch of short vertical videos. The software transcribes the source, picks moments it expects to hold attention, adds burned-in captions and reframes the speaker for 9:16 feeds such as Shorts, Reels and TikTok. Some tools then schedule the posts or expose an API so the whole route can run inside an automated workflow.

We examine what happens after the demo: how many clips are actually publishable, how long repairs take, whether captions get names right, and what each accepted clip really costs. We do not rank tools on feature counts, vendor “virality” scores or affiliate payouts, and we do not review general video editors or AI avatar generators.

Diagram of a clipping AI workflow in four steps: upload a long recording, the tool cuts captioned vertical clips, a person reviews and rejects weak clips, then approved clips are scheduled and posted.
Fig. 01 — The four stages we measure. Scrutator schematic, not product imagery.

Selection grid

Match the tool to the job.

Start from your situation. Each row shows what matters most before you compare any product.

Your situationClip controlAutomationCaption qualityCost structureAPI
Solo creator making volumeMedium — fix the odd cutHighGood enough to postLow monthly feeNot needed
Team reviewing before publishHigh — trim and reframeMediumBrand names correctSeats and shared creditsOptional
Automated pipelineLow — rules set upfrontVery highConsistent and checkableCredits per source minuteRequired, with recovery
Budget firstMediumMediumTolerates manual fixesCost per accepted clipNot needed

Keep digging

Clipping AI, examined from different angles.

View the directory ↗
01

Comparison framework

Opus Clip alternatives

A lower subscription price is only a saving if the resulting clips need less work.

↗
02

Test plan

Vugola vs Opus Clip

The same source, the same export target, and a cost comparison that counts usable clips.

↗
03

Published prices checked

Vugola AI pricing

Published monthly plans, credit arithmetic, and why a credit is not the same as a finished clip.

↗

New research files

Opus Clip, examined.

Documentation-checked files on Opus Clip review, pricing, the free plan, the API and the best AI clip makers, plus step-by-step Opus Clip guides.

Open the files ↗

What a Scrutator page covers and leaves out is set out once on our methodology page.

Our ground rule

The test comes
before the verdict.

Documentation is checked, and six tools have been tested hands-on: Opus Clip and Vugola on October 19, 2026, and Klap, Pictory, Descript and Captions on October 29–30. Scores are provisional — one tester, one round — and are not a stable industry ranking. We publish a score only when the run record can support it.

Read the testing protocol ↗

Questions before you start

FAQ

What is a clipping AI tool?

It is software that takes a long video, such as a podcast, webinar or interview, and proposes short vertical clips from it. Most tools pick moments, add captions and reframe the speaker for 9:16. You still decide which clips are worth publishing and fix the ones that are close.

Are these tools free?

Some offer a free tier or trial, usually with watermarks, export limits or short retention. The tools we cover so far sell monthly plans with a credit allowance. Check the live pricing page before uploading anything, and judge cost per clip you actually publish, not per credit.

How does Scrutator test these tools?

We run the same three rights-cleared sources through each tool twice: a tutorial, an interview and a screen demo. We log every clip as accepted, repairable or rejected, measure caption errors and repair minutes, and record what each job actually charged.

Are the scores on this site final?

No. Scores are provisional: one tester, one round per tool, and not a stable industry ranking. Opus Clip scored 73.5 and Vugola 51.6 on October 19, 2026 (12 jobs). Klap scored 82.4 and Pictory 79.0 on the same protocol on October 29. Descript’s 76.8 is a partial sample; Captions has clip grades but no overall score. The full table is on our hands-on test results page. Tools we have not run stay unscored.

Which tool should I start with?

Start from your bottleneck, not a ranking. If you need an API or agent workflow, read the Vugola AI review. If you are leaving Opus Clip, use the alternatives worksheet. Then run one of your own videos through a trial before paying for a year.