Habr author compares 8 AI services for Reels and Shorts: editing time shrinks to minutes

One to two hours of recorded material used to eat up one or two evenings of editing. Now, according to the author of a recent review on Habr, the same volume turns into a set of vertical clips in a few minutes. He tried quite a few free and paid services for cutting Reels and Shorts, and kept eight in his final selection that, in his assessment, genuinely work in 2026.

The published excerpt does not list the services; the names are hidden under the cut. So let's look at what the claim itself reveals: what work the AI takes over, where such tools tend to stumble, and how to test them before paying for a subscription.

What exactly gets automated

Cutting a long video into short clips by hand involves several steps: watching the recording, finding complete thoughts, cutting them out, reframing the horizontal picture to vertical, adding subtitles, and exporting. The first step is the most labor-intensive. You have to sit through the entire recording and make a hundred small decisions about what will hook a viewer.

That is the part the new generation of services takes over. The typical pipeline works like this: speech is transcribed, a language model marks fragments that make sense on their own, an algorithm reframes the picture to 9:16 and keeps the speaker in shot, and subtitles are overlaid on top. The human is left to review the finished options and pick the best ones. That explains the gap between an evening and a few minutes: the machine does the rough work, and the human does the editing.

Key figures from the review

• Input material: 1–2 hours of footage.
• Manual editing: one or two evenings.
• With an AI service: a few minutes, by the author's estimate.
• Scope of the review: 8 services, both free and paid.
• Output format: 9:16 vertical video for Reels and Shorts.

The figures come from the author and are not backed by independent measurements. Treat them as personal experience, not as a benchmark.

Where such tools usually break

The category has typical weak spots, and those are what you should compare services on.

First: choosing moments. A model can cut out a striking phrase but strip it of its context, and the clip turns into a fragment. Second: the Russian language. Transcription often confuses names, terms, and anglicisms, and fixing subtitles by hand eats into some of the time saved. Third: framing. When two or more people are in shot, or the speaker moves a lot, automatic cropping often loses the subject. Fourth: pricing terms. Free plans typically come with watermarks, limits on minutes of source footage, and caps on export resolution.

These are not complaints about any specific product in the review, but a list of questions it makes sense to ask of any of them.

How to choose in practice

Start with your own recording, not with demo clips. Take the same hour-long recording and run it through two or three services. Then compare them on a few simple criteria:

• how many minutes passed between upload and finished clips;
• how many of ten suggested clips can be published without edits, and how many need redoing;
• how the subtitles handled your vocabulary and names;
• whether the framing kept the speaker in shot;
• what exactly sits behind the paid subscription: resolution, watermark, source volume.

Check the terms on rights to the material separately if the clips include someone else's speech or music. And don't publish the result blindly: a five-minute review of the selected clips remains the cheapest way to avoid losing your audience to a clumsy cut. If a service saves you an evening but demands an hour of fixes, the gain is smaller than the advertising promises.

The full list of eight services and the author's impressions are available in the original article on Habr.