Hi Product Hunt, maker of shortshort here. I built it because every clipping tool I tried on long talks and podcasts gave me the same three problems: clips that stopped mid-sentence, a 9:16 crop that drifted off the speaker, and captions one word late. So shortshort focuses on exactly those three things: 1. The AI watches the whole video (picture and audio), picks distinct passages and ends each one on a complete sentence. A short can run 25% shorter or longer than your target so a thought is not cut in half. 2. A face tracker moves the 9:16 crop with the speaker, keeps camera cuts as hard cuts, bounds zooms between x1.0 and x1.6, and leaves slides and wide shots full-frame. 3. Whisper times every word, so captions follow the voice in one of four styles. A 45-minute podcast episode or a conference keynote comes back as a series of MP4s in 1080x1920 (3 on the free tier, up to 20 on Studio) while you close the tab. Every cut and keyframe stays editable, and each exported short gets a public before/after page with the source video and transcript. Pricing is by the minute: 1 credit = 1 minute analysed, whatever the number of shorts. 60 credits are free, no card. One question for you: when you clip a podcast or a talk today, what do you fix by hand most often, the ending, the framing or the captions? That is what I want to improve next.

Shortshort

One long video in. Up to 20 shorts, cut on full sentences.

One long video in. Up to 20 shorts, cut on full sentences.