Image to video · start + end frame · zoom out or zoom in
Earth zoom out AI — from your photo to outer space in one shot
The viral shot that starts on you and pulls back — over the rooftops, through the clouds, out to the whole planet. Make it from one photo, and add a picture of Earth as the end frame when you want to choose exactly where it lands.
- From one photo
- Optional Earth end frame
- Zoom out or zoom in
- Up to 30 s and 4K
Updated by the Kenerate AI team
On this page
Quick answer
How do you make an Earth zoom out video with AI?
An Earth zoom out video starts close on a person or object and pulls the camera back in one continuous shot — over the city, through the clouds, out to the whole planet. To make one, upload your photo as the start frame of an image-to-video model and describe the pull-back to space. For a precise ending, add a picture of Earth from space as the end frame: on Kenerate, Seedance 2.5, Seedance 2.0, Kling v3 Pro, Veo 3.1 and Wan 3.0 take both. Reverse the clip for the zoom in version.
Key facts
- Effect
- One continuous pull-back from you to the whole Earth
- Input
- Your photo as the start frame; an Earth image as the end frame (optional)
- End-frame models
- Seedance 2.5, Seedance 2.0, Kling v3 Pro, Veo 3.1, Wan 3.0
- Start frame only
- Sora 2 Pro, Pixverse v6, Happy Horse 1.1
- Length
- 2–30 s, by model (5–10 s is plenty)
- Resolution
- Up to 4K on Seedance 2.5 and Veo 3.1
- Zoom in
- Swap the frames, or reverse the clip
- Plan
- Video uses credits from a credit pack
Methods
Three ways to make the Earth zoom, compared
All three run in Video Studio's image-to-video mode. The end frame is what decides how the shot finishes.
| Method | What you upload | What you control | Models on Kenerate |
|---|---|---|---|
| Photo + zoom-out prompt | Your photo as the start frame | The start; the model invents the journey and the final Earth shot | Any image-to-video model |
| Start + end frame | Your photo, plus a picture of Earth from space as the end frame | Both ends; the model fills in the pull-back between them | Seedance 2.5, Seedance 2.0, Kling v3 Pro, Veo 3.1, Wan 3.0 |
| Zoom in (frames swapped) | The Earth picture as the start frame, your photo as the end frame | A dive from space down to you, landing on your exact photo | The same end-frame models — or reverse a zoom-out clip in any editor |
Prompts
Earth zoom prompts to copy
Each one opens Video Studio in image-to-video mode on Seedance 2.5 with the prompt filled in. Add your photo as the start frame.
- Subject
- Who or what the shot starts on
- Camera
- Pulls back fast, rising straight up
- The way up
- Street, city, coast, clouds, space
- Ending
- The whole Earth — or your end frame
Classic pull-back from a person
Start on the person in this photo. The camera pulls back fast and smoothly, rising straight up: the street shrinks into the city, the city into the coastline, the camera climbs through white clouds and the blue sky into black space, ending on the whole Earth slowly turning. One continuous shot, no cuts; the person stays the same.
From an object
Start close on the coffee cup on the café table in this photo. Fast, seamless zoom out: the table, the café, the street outside, the whole city from above, then up through the clouds into space, ending on planet Earth against black space. One continuous camera move, no cuts.
Travel photo
Start on the person on the beach in this photo. The camera rises and pulls back: the beach, the bay, the whole coastline and island, then through the atmosphere to orbit, ending on the Earth with the island just visible. Smooth, accelerating, one continuous shot.
With an Earth end frame
One seamless continuous zoom out from the first frame to the last frame: the camera pulls straight up and back from the person, over the rooftops and the city, through the clouds, and arrives exactly on the Earth seen from space in the last frame. No cuts, no fades.
More prompts (2)
Earth zoom in (frames swapped)
One seamless continuous zoom in from the first frame to the last frame: the camera dives from the Earth seen from space, down through the clouds, over the city and the street, and lands exactly on the person in the last frame. No cuts, no fades.
Vertical reel
Vertical 9:16. Start on the person in this photo looking up at the camera. Rapid upward pull-back: the park, the neighbourhood, the city lights, the curve of the planet, ending on the whole Earth glowing in space. Whoosh of wind as the camera rises, then silence in space.
How to
How to make an Earth zoom out video
- Time needed
- Steps
- 4 steps
Step 1
Pick your photo
A sharp, well-lit outdoor photo of yourself, a friend who agreed, a pet or an object, with the background visible. Avoid signs and small text.
Step 2
Get an Earth end frame (optional)
For a fixed ending, make a picture of the Earth from space in Image Studio, or use a NASA image of Earth with a credit line. Match its shape to your video (9:16 or 16:9).
Step 3
Open Video Studio
Press “Make an Earth zoom out”: Video Studio opens in image-to-video mode on Seedance 2.5 with the prompt filled in. Upload your photo as the start frame and, if you have one, the Earth picture in the End frame slot.
Step 4
Set length, shape and sound
5–10 seconds, 9:16 for Reels and TikTok or 16:9 for YouTube. Sound models add sound unless you turn it off.
See it in the app

Open Video Studio in Text to Video mode and write the shot — subject, action, camera move and light.
Watch the tour · 1:19
Text, image & reference to video, models, edit, extend & chat
Chapters
Tips
Prompt and photo tips for a clean zoom out
What keeps it one smooth shot instead of a jump cut to a stock globe.
Use a sharp outdoor photo with the surroundings visible — the model needs a city, coast or landscape to pull back through.
Write the path in stages: subject, street, city, coastline, clouds, space, the whole Earth.
Say “one continuous shot, no cuts” — without it some models cut straight to space.
Add an Earth end frame for a fixed ending; generate one in Image Studio, or use a NASA photo and credit NASA.
Keep it short: 5–10 seconds is enough. Longer clips tend to wander on the way up.
Draft at 480p or 720p, then remake the take you like at 1080p or 4K.
Models
Video models for the Earth zoom
Every model below takes your photo as the start frame and a picture of Earth as the end frame.

ByteDance
Seedance 2.5
- Reference images
- Start + end frame
- Output
- 4–30 s · up to 4K
Open Seedance 2.5Best for
Default here · long, cinematic camera moves

ByteDance
Seedance 2.0
- Reference images
- Start + end frame
- Output
- 4–15 s · up to 4K
Open Seedance 2.0Best for
Image-to-video with sound
Kuaishou
Kling v3 Pro

Google
Veo 3.1
- Reference images
- Start + end frame
- Output
- 4, 6 or 8 s · up to 4K
Open Veo 3.1Best for
Google's first-and-last-frame transitions
Alibaba
Wan 3.0
- Reference images
- Start + end frame
- Output
- 2–30 s · up to 1080p
Open Wan 3.0Best for
Cheap drafts to test the path
FAQ
Earth Zoom Out Video questions
10 questions
What is the Earth zoom out effect?
It's a viral AI video shot that starts close on a person or object and pulls the camera back in one continuous move — over the rooftops, the city and the coastline, through the clouds — until the whole planet fills the frame. Apps like Higgsfield and PixVerse offer it as a one-tap template; with an image-to-video model you write the path yourself.
Link to this answerHow do I make an Earth zoom out video with AI?
Upload a sharp outdoor photo as the start frame in an image-to-video model and describe the pull-back: subject, street, city, clouds, space, the whole Earth, “one continuous shot, no cuts”. On Kenerate, the button on this page opens Video Studio on Seedance 2.5 with that prompt filled in; add a picture of Earth as the end frame if you want to fix the final shot.
Link to this answerWhich AI video models support start and end frames?
On Kenerate, Seedance 2.5, Seedance 2.0, Kling v3 Pro, Veo 3.1 and Wan 3.0 take both a start frame and an end frame in image-to-video mode, along with several older models. Sora 2 Pro, Pixverse v6 and Happy Horse 1.1 take a start frame only. The studio shows “no end frame on this model” when a model can't take one.
Link to this answerDo I need an end frame for the Earth zoom?
No — one photo and a good prompt is enough, and the model will invent the final view of Earth. Add an end frame when the ending matters: a specific continent, a night-side view with city lights, or a picture that matches your brand. The model then fills in the journey between your two frames.
Link to this answerWhat prompt makes the Earth zoom out?
Describe it in four parts: the subject and setting, the camera move, what appears on the way, and the final shot. For example: “Start on the person in this photo. The camera pulls back fast and smoothly, rising straight up: the street, the city, the coastline, through white clouds into space, ending on the whole Earth slowly turning. One continuous shot, no cuts.”
Link to this answerHow do I make the Earth zoom in version?
Two ways. Swap the frames — the Earth picture as the start frame, your photo as the end frame — and prompt a dive from space down to the subject. Or make the zoom out and reverse the clip in any video editor. Swapping frames gives a cleaner landing on your exact photo.
Link to this answerIs the Earth zoom out AI free?
You can open Video Studio and set the shot up without paying, but generating video uses credits from a credit pack; a free account's starter credits are for images. Credits for the exact model, length and resolution are shown before you generate. Draft short and low-resolution first.
Link to this answerCan I use a NASA photo of Earth as the end frame?
NASA says its images are generally not copyrighted in the United States, so many NASA Earth photos can be used if you credit NASA. Don't use NASA logos, and don't suggest NASA endorses your video. Images NASA marks as third-party copyright need that owner's permission. Or generate an Earth picture yourself in Image Studio.
Link to this answerWhy does the face change during the zoom?
The model redraws every frame, and a face that shrinks to a few pixels and back can drift. Use a clear photo, keep the clip short, say “the person stays the same”, and for the zoom in, put your real photo in the end frame so the landing matches it.
Link to this answerCan I use photos of other people?
Use photos of yourself, of people who have agreed, and of places, pets and products you have the right to use. Don't make clips of real people without their consent, and don't present an AI zoom as real drone or satellite footage.
Link to this answerIn depth
Earth Zoom Out Video, explained
What works, what to avoid, and how to pick a model.
On this page
How the Earth zoom was made before AI
Editors built it by hand: a stack of aerial and satellite pictures from closest to farthest, each scaled up and faded into the next, so the camera seemed to rise in one move. An image-to-video model generates the in-between views itself, which is why one photo and a sentence now does the job.
Why the end frame matters
With a start frame only, the model decides what “the whole Earth” looks like — and some takes cut away or end on a generic globe. With an end frame, you choose the ending and the model only has to connect two fixed pictures. Google describes the same idea for Veo 3.1: give a first and a last frame, and it generates the transition.
What can go wrong
Common failures, and the fix for each:
- The shot cuts to space halfway: add “one continuous shot, no cuts, no fades”.
- The city looks nothing like your photo's place: name it — “a coastal town”, “a desert highway”.
- The zoom stalls in the clouds: shorten the clip, or add an end frame.
- Text on signs turns to gibberish: pick a photo without signs or logos.
Make your Earth zoom out video
Set up the shot for free, see the credits before you generate.
Tried it? Tell us how it went





