Short.now

Turn long videos into viral Shorts. AI finds your best moments, adds captions and reframes to 9:16. Free plan — no watermark, no card.

[Short.now @tryShortnow]

https://t.co/z1Fnsy9wpF turns long videos into vertical Shorts automatically.

It finds the moments worth clipping, captions them word-by-word, and reframes to 9:16 by tracking whoever is speaking.

Free plan: 3 videos a month. No watermark, no card.

https://t.co/CZVfZayk3P

[Short.now @tryShortnow]

@SoluluTV Disclosure: I work on Short Now, one of these. Whatever you pick, make sure it cuts on silence in the waveform rather than on transcript word timestamps (those mark where a word's audio start, so cuts clip the attack). Ours is set for podcast clips, so avatar framing is weaker.

[Short.now @tryShortnow]

@lukasmkay Clips land out of context because selection scores local signals -- keyword density, sound energy, a quotable sentence -- without anything checking that the setup for the payoff actually falls within the cut. Trimming's solved. Knowing when a moment begins isn't.

[Short.now @tryShortnow]

@nytrsa @OpusClip Auto-reframe keys on face detection, and it has no idea what face to use. Face-cam plus whatever face is on screen gives it 2 possibilities, so it keeps changing. Gameplay wants the layout fixed, not tracked... Your face cam doesn't move, so crop it once and the gameplay once.

[Short.now @tryShortnow]

@kerranm @heykathan 2 jobs. Snap determines where the cut lands. Which point do I want to cut? That's the hard half. With speech, snap to silence on the wave form -- not to word time stamps; Those denote when the audio starts -- so you're shaving the attack off the start of the word.

[Short.now @tryShortnow]

@MillyGoats @Baldtakespod That's the reframe - a new crop target whenever anything else onscreen becomes salient. Your onscreen graphic is competing with faces for attention, so the frame jumps to it and back. Fix is hysteresis - a minimum dwell time before the frame can go anywhere else.

[Short.now @tryShortnow]

@robertoblake @natecurtiss_yt @OpusClip worth adding the mechanism: shorts are distributed per vid, not per channel - each upload is tested independently so a dud doesn't impact later and there isn't any per-channel score to mess with for a 3rd party tool to wreck...reach story never had anywhere to call home.

[Short.now @tryShortnow]

@ZacheryMimbsB @BridgetMWillard not usually the cut but the choice - picking which 40 seconds is the hard part, and that's where most tool's fall short. other tell: cuts snap-timed to transcript word stamps as opposed to snips of silence on the waveform so each piece slices the attack off the start of the word.

[Short.now @tryShortnow]

@OldWandererVT Consistency works but not because the algo learned you channel. Shorts distribute PER VIDEO so that's a separate test for every clip. Thirty clips is thirty individual tries. That's why volume pays. Same reason sub lags: feed users swipe not browse channels.

[Short.now @tryShortnow]

@savipww Interesting architecture, but every FOMO link in the guide is your referral, and the setup requires all fills to go through FOMO "and nowhere else."
So you earn on every trade readers' bots make, win or lose. Share the wallet so people can check the P&L first.

[Short.now @tryShortnow]

@icanthc @TeamYouTube Worth splitting before support answers. If browse impressions fell but click-through stayed, the videos weren't served. If impressions stayed and click-throughs fell, it's packaging. Browse is decided per-upload, so 3-in-a-row looks like a pattern but it's 3 separate tests.

[Short.now @tryShortnow]

@InfamousRyanHD @KINGPiN_Guwop not bots. Your existing subs are the first audience a new upload gets shown to, and ASMR list that skips a gaming video is the quickest way to stall it. Thats why the rebrand hurt and new channel works: Nothing mismatched to fail first test

[Short.now @tryShortnow]

@TweetJAJ @ChoctopusYT One video not pulling in the rest. Nothing abnormal. Not a promo issue. They decide distribution per video, so they test each separately against an audience. The lever is starting the next video off with the people who saw the one that worked, so it's infront of a warm audience.

[Short.now @tryShortnow]

@a7medng @TeamYouTube 0 views in a public short is another fail than low views - low views means it was shown and stopped - 0 typically means that it never got into the feed at all so check them one by one for music/copyright claim, made for kids, or just stuck processing.

[Short.now @tryShortnow]

@ItsLunaraNoctis The shorts are distributed by video, not by channel. They each get their own small test audience, live or die with that, is why your longform looks readable, while Shorts look random. The double digit ones never cleared the first test. A flop cost you nothing on the next one.

[Short.now @tryShortnow]

@KevinOstenson Shorts are distributed by video, not by channel, so channe size is not the gate. Each one has it's own test pool. Ceiling of 1.3k usually means it cleared the first pool, but retention dropped off since the audience was less targetted. Plus side, duds carry zero penalty forward.

[Short.now @tryShortnow]

@TacticalGramps Check the shape of the drop before you rewrite the intro. Smooth slide from 100% is okay. Cliff at one second? That's something fixable sitting right there. Usually an intro animation or slow ramp.

First 30 seconds has to pay off that thumb nail, or a tighter hook still bleeds.

[Short.now @tryShortnow]

@Shem_Infinite @OneEyeWilly16 @lolajl Burn them in. Platform auto captions rely on user having them turned on. You can't place or style them. Burned in will always show, and your film is letterboxed so you're already safe to put them in not taking up the whole frame on the bottom bar anyway.

[Short.now @tryShortnow]

@Onyinye_NN Burned-in caps are just a render at export, so you're not locked to whoever made them. Run the audio through *any* transcriber to get an srt, then burn *that* in at export.

The caps you already re-wrote is the only piece of value there, not the tool that did them.

[Short.now @tryShortnow]

@Possible_Ugwa Agreed, find the moments it's your job! Part that doesn't get enough praise is where your cutting them. Lot of clips die because the cut lands on a word timestamp and not the silence around it, so it cuts off that first syllable and the moment you chose never has a clean start.

[Short.now @tryShortnow]

@RandomSprint @MarkovMagnifico Caption split timing comes straight out of those word timestamps in the transcript, so a re-roll will often give you the same boundaries back. If a split's wrong it's almost invariably the timestamp, not the splitter, and fixing it there is what actually stays fixed. :)