AI video clipper for long videos

What an AI video clipper actually does

An AI video clipper turns one long recording into several short, vertical videos that stand on their own. Instead of scrubbing through an hour of footage for the forty seconds worth posting, you hand the clipper the file or a YouTube link and it does the watching: it transcribes every word, finds the moments with a clear hook and a payoff, trims them to a length that suits Shorts, Reels and TikTok, reframes the 16:9 shot to 9:16 and adds captions

The hard part is not cutting video, it is choosing. A good clip opens on a line that makes someone stop scrolling, makes its point without the context around it, and ends before attention drops. shitShorts scores every candidate window for exactly those things, so the shortlist you see is already filtered

How shitShorts picks the moments

First the audio is transcribed with word-level timestamps, so every cut can land between words rather than in the middle of one. The transcript is split into overlapping windows and each window is scored by a language model for hook strength, payoff, pacing and whether it makes sense on its own. A second, stricter model then reviews the top candidates, tightens their start and end points and drops the ones that only work with the surrounding conversation

Signals from the footage feed the same score: laughter, applause and audio peaks, scene cuts, and how many faces are on screen. A strong opinion delivered on camera with a reaction beats a strong opinion read off a slide. Each clip arrives with its reasons, so you can see why it made the cut

Reframing to vertical without cropping people out

Most long-form video is shot wide, and a centre crop to 9:16 cuts off whoever is not in the middle. The clipper tracks every face through the clip, matches voices to faces with speaker diarization, and moves the vertical frame to whoever is talking. When two people share the moment it switches to a split layout instead of a crop that shows neither of them well. On sports footage it follows the action rather than faces

Captions, hooks and edits that stay editable

Captions are generated from the same word-level transcript and placed inside the safe area of each platform, above the like and comment buttons. You can switch the caption theme in one click without re-timing anything

Every clip opens as a normal project in the browser Studio: the crop, captions, hook text and zoom-cuts are separate layers you can move, restyle or delete. Nothing is burned in until you export, so fixing a misheard name takes seconds instead of a full re-render

Who it is for

Podcasters and interview shows who want a week of Shorts from each episode, streamers turning VODs into highlights, coaches and educators clipping talks and webinars, sports channels cutting highlight reels, and agencies doing all of the above for clients. If the video has people talking or clear action, it is a good fit

What it costs

The free plan includes 300 credits every month, roughly 15 clips, with 720p exports and a small watermark. Paid plans remove the watermark: the Clipper plan is $10 a month for 1,000 credits, about 50 clips, at 1080p with 300 clip exports a month, while Creator ($19) and Pro ($49) add more credits, 1080p60 or 4K exports and the scheduler. On the free plan the first 3 clip exports are free and each one after that costs 10 credits. You see the credit estimate before a run starts, a failed run is refunded automatically, and editing, captions and previews never cost credits