JUMPon theTREND
The Dolly Zoom AI trend in short
What it is: You stand still on a long dusk street as the camera glides in: you stay the same size while the street behind you stretches away.
How to do it: Add a clear selfie on this page, see yourself in it free, then make the 5-second video. No app, no prompt.
Who to put in it: You, The friend who always realises a second too late, The travel buddy who wants the mountain shot without the hike.
Dolly Zoom: posts and how-tos
- Karen X Cheng · an AI-made dolly zoom from phone footage, YouTube Short, 26 Jan 2023, 4.48M views (read 18 Oct 2026)
- Tanya Khanijow · a phone tutorial for pulling a mountain closer with a dolly zoom, YouTube Short, 8 Jul 2026, 845,944 views
- Now You See It · a video essay on the dolly zoom in film, YouTube, 24 May 2017, 2.8M views
- Starrd · the dolly zoom (vertigo) as an AI camera move, with its prompt and three templates that use it (sitemap date 16 Jul 2026)
- getimg.ai · a guide to its one-photo dolly zoom action, with three directions and five intensities, 29 Sep 2026
- Tripsy Pooja · a travel creator trying the dolly zoom as "this cool travel trend", YouTube Short, 14 Oct 2026, 26,873 views by 18 Oct
Clips belong to their creators and link to the original post. Creators can ask us to remove a link.
What is the Dolly Zoom AI trend?
The dolly zoom is the camera trick where a person stays exactly the same size in the frame while the world behind them stretches away or rushes in, as if the ground just shifted under them. Here you make it from one photo: you get a free preview still of yourself on a long dusk street in about a minute, then the 5-second video in 1080p for 100 credits ($9.90). The camera glides in while the view widens, so your head and shoulders never move in the frame, the street behind you races off into the distance and your calm face turns to wide-eyed realisation. People also call it the vertigo effect, the vertigo shot, the zolly or the trombone shot.
It's a film technique, not a single post. Irmin Roberts, a second-unit cameraman at Paramount, worked it out for Alfred Hitchcock's Vertigo, which premiered in San Francisco on 9 May 1958: to look down a bell-tower stairwell, the crew laid a one-tenth scale model of the stairs on its side, then tracked the camera back while zooming in (AFI Catalog). Far Out Magazine traces the idea back to Hitchcock's Rebecca in 1940, when the cameras of the time couldn't manage it, and Wikipedia notes that Roberts was never properly credited for the shot. Jaws used it in 1975 on Chief Brody's face at the beach. On phones it became a trend of its own: a compilation of TikTok vertigo-effect clips went up on YouTube on 1 Feb 2020, and Abby Roberts posted "dolly zoom trend back again" in September 2022. The first viral AI-made one we found is Karen X Cheng's, made from phone footage with a 3D-capture app and posted on 5 Dec 2022; her YouTube Short "NeRF Dollyzoom effect" had 4.48M views when we read it on 18 Oct 2026. By January 2026 travel creators were using it to pull mountains closer, and AI template sites list it too: Starrd's camera-move pages and media.io's effect page carry sitemap dates of 16 Jul and 29 Aug 2026, and getimg.ai's guide to its one-photo Dolly zoom action is dated 29 Sep.
Ours comes in two versions. The realisation is the classic: a long, empty old-town street at blue hour, lamps lit, running to a point right behind your head; the camera moves in as the view widens, the townhouses lean out and the far end runs away, and the clip holds on your face at the most stretched moment. Mountain pull is the travel version going round Reels and Shorts in 2026: you stand at a lookout, the camera pulls back while the view tightens, and a snowy peak swells up behind you while you haul it in, hand over hand, on a rope that isn't there. Your face, hair, facial hair, glasses and clothes come from your photo either way. It isn't a re-creation of any film scene: no stairwell, no bell tower, no beach. And because it starts from a photo, it can't add the effect to a video you've filmed yourself.
Is the Dolly Zoom trend still growing?
Evergreen. The dolly zoom has been a film staple since 1958 and a phone trend in waves since at least 2020. The newest wave is travel creators pulling mountains closer: a Hindi tutorial called the Mountain Zoom Effect viral on 2 Jan 2026, and Tanya Khanijow's tutorial of 8 Jul 2026 had 845,944 views on 18 Oct. Small AI-filmmaking channels have been testing whether video models can hold the move (June to September 2026). Search is steady rather than spiking: on 18 Oct Google completes "dolly zoom" ten ways, mostly film and how-to (jaws, vertigo, iphone, meaning), and "dolly zoom ai" four ways. Two of the template sites we track have it (Starrd, sitemap date 16 Jul; media.io, 29 Aug), Vidu sells an undated one-photo version, and our radar doesn't score it.
How to do the Dolly Zoom trend: by hand or in one click
By hand: put your photo into the scene with an image model, regenerate until the face looks like you, give that still to an image-to-video model with the moves written out, then check the face holds to the last frame.
Here, in one click: add a photo for each person, see a free preview still of you in the scene, then make the 1080p video from it. You pay only when it’s ready.
Dolly Zoom mixes
- The realisation
- Mountain pull: instead of the street, a railed mountain lookout in bright morning light, a huge snow-capped peak far behind across a green valley. The move runs the other way: the camera pulls back as the view tightens, so they stay the same size while the peak swells up behind them until it fills the top half of the frame; meanwhile they pull hand over hand on an invisible rope as if hauling the mountain in, then grin at the lens on the hold. Wind and a rising whoosh instead of the low tone; no visible rope or other people
Photo tips for Dolly Zoom
- One clear photo of you facing the camera, waist up or closer, well lit, with nobody else in it.
- Your face stays the same size the whole way through, so a sharp, front-facing photo matters most.
- Glasses stay if you wear them in the photo; take a hat or sunglasses off unless you want them in the shot.
- Wear what you want to be seen in: the clothes in your photo come with you onto the street or the mountain.
Who to cast
- You
- The friend who always realises a second too late
- The travel buddy who wants the mountain shot without the hike
Dolly Zoom FAQ
Is there a free Dolly Zoom AI video generator?
You can preview yourself in the Dolly Zoom scene free. The full 1080p video costs 100 credits, charged only when it's ready. Credits come in packs from $9.90 for 100 and work on any trend.
Which app makes the Dolly Zoom AI video?
Creators put their photo into the scene with an image model, then animate it with an image-to-video model. This page does both for you from one photo: a free preview still first, then the video.
How many photos do I need for Dolly Zoom?
One clear, front-facing photo per person (or pet). This template has room for 1. Add up to 3 photos of each for a closer likeness.
What happens to my photos?
They are stored privately, sent to the video model through short-lived links, never used for training, and deleted with your video after 30 days.
What is the dolly zoom AI video?
A short AI clip of the dolly zoom made from one photo: you stay exactly the same size in the frame while the background stretches away or looms closer. Here it's you on a long dusk street as it dawns on you, or at a mountain lookout as a peak swells up behind you.
Is the dolly zoom AI free?
The preview is free here: add a photo and see a still of yourself in the scene in about a minute, with no card. The 5-second video is 100 credits ($9.90), in 1080p. Plenty of pages promise a free dolly zoom effect online; we haven't checked what each one charges for the video.
How do I do the dolly zoom trend?
Here: add one clear photo, pick The realisation or Mountain pull, check the free preview, then make the video. To film it yourself, walk toward your subject while zooming out, or back while zooming in, at the same rate so they stay the same size; phone tutorials start on the 0.5x lens.
What's the dolly zoom AI prompt?
The prompts that work name the move both ways: the camera travels one way while the lens zooms the other, the subject stays the same size, and the background has strong lines to stretch. Here you don't write one: you add a photo and the scene, the move and the held ending are already set.
Why is it called the vertigo effect?
Because it was first used in Alfred Hitchcock's Vertigo, released in 1958, to show a detective's fear of heights as he looks down a bell-tower stairwell. Paramount cameraman Irmin Roberts worked out how to do it. It's also called the Hitchcock shot, the Jaws effect, the zolly and the trombone shot.
Which way does the background move?
Either way. Camera in while the view widens: the background stretches away, as in The realisation. Camera back while the view tightens: the background looms closer, as in Mountain pull. You stay the same size both ways.
How long is the video?
5 seconds, vertical 9:16, in 1080p: half a second still, about three and a half seconds of move, then a held beat at the most stretched point. getimg.ai's version runs 5, 6 or 10 seconds; Starrd and media.io don't say.
Can I add the effect to my own video or drone footage?
Not here: ours is made from one photo. For your own footage you'd film it that way (some drones have a dolly zoom mode) or fake it in an editor with a keyframed zoom, which plenty of tutorials show.
Is it the same as bullet time?
No. In Bullet Time the moment freezes while the camera travels round you; in a dolly zoom time runs normally, the camera moves straight in or out, and the lens zooms the opposite way.
Can I do it with my dog or cat, or with a friend?
Not on this one: it's one person, waist up. For your pet, see the AI pet trends; for two of you, the Korean baseball fan cam has a two-of-you version.
What song is used for the dolly zoom trend?
There isn't one. Dolly zoom posts use whatever sound fits the place. Yours comes with a low swelling tone (or wind on the mountain) and no music, so add a track in TikTok, Reels or Shorts when you post.
What photo works best?
One clear, well-lit photo of you facing the camera, waist up or closer, with nobody else in it and no hat or sunglasses. Your face stays the same size the whole way, so sharpness matters more than the background, and your clothes come with you.
Is my photo kept?
For 30 days, then it's deleted automatically, and it's never used to train AI. You can delete everything sooner with Delete my data on Your videos; the details are in our safety guide.
More AI trends like Dolly Zoom
Your photo carries over. Preview yourself in any of these.
Make yours
Make yours
Make yours
Make yours
Make yours
Make yours
See every AI trend this week Not here? Suggest a trend → · Video templates → · Best photo → · How to do it → · Is it safe? →
Independent tool. Not affiliated with TikTok, the creators of the trend posts, or the rights holders of the songs. Only upload people who agreed to be in it.