AI Clip Maker: Turn Long Videos Into Shorts
BetaAuto Clips is a free AI clip maker that runs in your browser. It transcribes a long video, scores every stretch of speech for pace, distinctive words and energy, and cuts the best moments into 15, 30 or 60-second clips, framed around the speaker and captioned. The video never leaves your device. It is in Beta: treat the picks as a strong first draft.
- Runs on your device
- Whisper AI
- On-device AI
- Works offline
Updated
This tool is in beta: it works, but it is heavy for some devices and results may need a second look.
Drop your video here
How to turn a long video into shorts
Step 1: Drop a podcast, stream, interview or lecture video on the box above.
Step 2: Choose the Clip length (15, 30 or 60 s), the Number of clips (up to 10) and the Shape: 9:16, 1:1 or 16:9.
Step 3: Keep Add captions on for burned-in captions, then press Create and confirm the one-time speech model download.
Step 4: Preview each clip. Its card says where it came from and why it was picked; download the ones you like.
Good to know
Settings explained
- Spoken language
- Picking the language skips detection and improves accuracy on short clips.
Formats
How it runs
- Whisper speech model
- Speech recognition uses an open Whisper model. It downloads once after you confirm its size, then runs on your device — on the GPU via WebGPU when available.
- Vision model
- A compact computer-vision model runs locally in your browser. Nothing is sent to an AI service.
- Hardware video engine
- Video is decoded and encoded by your device’s media hardware through WebCodecs — usually several times faster than real time.
How the AI clip maker picks moments
Clips start and end where a transcript segment starts or ends, so they never cut into a word, and two clips never overlap. Each candidate stretch gets a score from five signals, and the highest-scoring ones become your clips. The result card shows the reason, such as “About budget planning” or “Dense, energetic speech”.
| Signal | What it measures | Weight |
|---|---|---|
| Pace | Words per second: dense talking beats rambling | 0.45 × words per second |
| Distinctive words | Words frequent in this stretch but rare elsewhere in the video | 0.35 × salience |
| Energy | Loudness compared with the whole video, capped at ±3 dB | 0.12 × dB difference |
| Hook | Opens with a question, an exclamation or words like how, why, secret, mistake, tip | +0.4 |
| Clean ending | Ends on a full sentence | +0.25 |
What is AI clipping?
AI clipping is finding the most shareable moments of a long video automatically and cutting them into short vertical videos for TikTok, Reels and Shorts. Cloud tools do it on their servers after you upload. This clipper, an Opus Clip alternative without the upload, does every step in your browser: speech recognition, scoring, face detection, cropping, captions and encoding.
Clip size and framing
Each clip is cropped from the original frame, not upscaled, so its size depends on your source. Whenever the shape needs a crop, it is centred on the speaker’s face, found on four frames of that clip; if no face is found it is centred on the frame. The crop stays still for the whole clip.
| Source video | 9:16 clip | 1:1 clip | 16:9 clip |
|---|---|---|---|
| 1280×720 (720p) | 406×720 | 720×720 | 1280×720 (no crop) |
| 1920×1080 (1080p) | 608×1080 | 1080×1080 | 1920×1080 (no crop) |
| 3840×2160 (4K) | 1216×2160 | 2160×2160 | 3840×2160 (no crop) |
| 1080×1920 (vertical) | 1080×1920 (no crop) | 1080×1080 | 1080×608 |
Turn a long video into shorts for free
Every clip length fits the platforms’ limits: YouTube Shorts take up to 3 minutes, Instagram Reels 3 minutes and standard X accounts 140 seconds. Clips are MP4 files with sound, so they upload as they are.
With captions on, words appear in the Pop style, three at a time in 9:16 clips and five in square or wide ones, with the spoken word highlighted in yellow. For podcast clips on Instagram, choose 9:16, 30 seconds and three to five clips, then post the ones that work on their own without context.
What the Beta can and cannot do
- Speech drives everything. A video without talking, such as music or gameplay without commentary, has no moments to score; use split video instead.
- One framing per clip. The crop does not follow someone who walks across the frame; auto reframe video tracks faces throughout.
- The picks are a draft. Trim any clip further with the video trimmer.
- Long videos need a computer. A recent laptop, ideally with WebGPU, handles an hour-long podcast; on a phone, use shorter sources.
- Private and free. No upload, no watermark, no clip credits.
Tips & troubleshooting
Want different captions? Turn Add captions off, then style each clip in the subtitle generator with your own font and colours.
Need a title and hashtags for each clip? Run it through the YouTube description generator, which also writes TikTok and Instagram captions.
Clips too quiet next to other posts? Even them out with normalize audio before uploading.
Ask for more clips than you need and keep the best: the video is transcribed once, so extra clips only add rendering time.
Auto Clips (Long Video to Shorts) FAQ
How do I turn a long video into shorts?
Drop the video here, pick 9:16, a clip length and how many clips you want, and press Create. Auto Clips transcribes it, scores every stretch of speech, and renders the top moments as vertical MP4 clips with captions, ready for Shorts, Reels or TikTok.
Is there a free AI clip maker without a watermark?
Yes, this one. It runs entirely in your browser, so there are no credits, no export limit and no watermark. The only large download is the speech model, once, after you confirm its size; the video itself is never uploaded.
Why is Auto Clips in Beta?
The highlight picks come from small on-device models and simple signals (pace, distinctive words, energy), not from a large cloud model, so they are suggestions rather than a finished edit. The crop is also fixed per clip. Review each clip and trim where needed.
How do I make podcast clips for Instagram?
Drop the video podcast, choose 9:16, 30 or 60 seconds and Add captions, then press Create. Each clip is framed on the speaker’s face and captioned in the Pop style. For an audio-only podcast, make a captioned video with the subtitle generator first.
Does it need a powerful computer?
The speech model and video encoding do the heavy work. A recent laptop handles long videos well, especially with WebGPU in Chrome or Edge. Phones work for shorter sources; for a two-hour stream, use a computer.