ShortsAIAn AI clip maker that picks the moments for you
ShortsAI is an AI clip maker: paste a link or upload a file, and it returns 3 to 15 clips of 15 to 60 seconds, each reframed to vertical and subtitled.
The clipping is automatic. Google Gemini 3.1 Flash-Lite scores the transcript and scene data, so you approve clips rather than scrub a timeline.
It is completely free: no paid tier, no watermark, no usage cap at shortsai.app.
What does an AI clip maker do that a video cutter does not?
A video cutter trims where you tell it to. An AI clip maker decides where to cut. That is the whole difference, and it is the difference that saves the time: a 60-minute podcast has perhaps a dozen moments that work on their own, and finding them by scrubbing takes longer than the recording did.
ShortsAI does the finding in three passes. faster-whisper transcribes the audio with a timestamp on every word. PySceneDetect marks the visual cuts. Google Gemini 3.1 Flash-Lite then reads the timestamped transcript together with the scene boundaries and returns the strongest 3 to 15 segments, scored on hook strength, emotional payload and whether the segment makes sense without what came before it. Giving the model the scene boundaries is what stops a clip from starting mid-shot.
How do you make clips with it?
- Paste a YouTube link or upload an MP4 or MOV. Podcasts, webinars, livestream recordings, interviews and vlogs all work.
- Pick the output format. 9:16 for Shorts, Reels and TikTok; 1:1 for feed posts; 16:9 if you want to keep the landscape frame.
- Wait for processing. On a typical CPU an 8-minute video takes 5 to 8 minutes; on an NVIDIA GPU about 50 seconds.
- Review the clips. Each one has a title, a hook overlay and burned-in subtitles, and you can re-cut, change the subtitle style or dub it into another language.
- Download, or publish straight to TikTok, Instagram Reels and YouTube Shorts.
How does the auto clipper handle vertical cropping?
Two modes. TRACK follows one subject with MediaPipe face detection and a YOLOv8 fallback, damped so the crop holds still inside a safe zone instead of chasing every head movement. GENERAL keeps the full width over a blurred backdrop for group shots and landscapes. A speaker tracker stops the crop from flipping between two people mid-sentence, and the split layout stacks both speakers when a conversation needs it.
Is the AI clip maker free?
Nothing. ShortsAI is completely free with no paid tier, no watermark and no usage cap, at shortsai.app. You bring your own Google Gemini API key, whose free tier covers 1,500 requests a day; the site is supported by ads.
It is not a trial. There is no metering code in it, which is why there is no cap and no watermark. What you bring is a Google Gemini API key, whose free tier covers 1,500 requests a day.
What are the limits?
The moment detection is only as good as the transcript, so a video with no speech (a music performance, a silent tutorial) gives it little to score. Clips are 15 to 60 seconds by design; if you need 3-minute cuts, the clip editor lets you set boundaries by hand. Caption styling is functional rather than lavish: the commercial tools ship more presets.
Common questions
- Can the AI clip maker work from a YouTube link?
- Yes. Paste the URL and the pipeline downloads the source, transcribes it and clips it. Local MP4 and MOV uploads work too.
- How many clips does it make from one video?
- Between 3 and 15, each 15 to 60 seconds. The count depends on how many moments actually stand alone, not on a quota.
- Does it add subtitles automatically?
- Yes. Subtitles come from the same word-level transcript and are burned in with FFmpeg, so they land on the right frame.
- Does it work without an NVIDIA GPU?
- Yes. A GPU is optional and cuts processing time by roughly an order of magnitude, but the CPU path works on any machine with Docker and 8GB of RAM.