
How do you turn an image into a video with AI?
By Alverson Layne, Founder
Founder of FXI Studio, building the AI film studio.
Published October 6, 2026 · Updated October 6, 2026
The short answer
Attach the image as the first frame of a video model and write a prompt that describes only the motion: what moves, how the camera moves, and where the shot ends. Add a last frame when the shot has to land on a specific picture. Generate a short clip, check it, and adjust the motion words rather than the scene.
TL;DR
The image already decides the look, so the prompt should only decide the motion. In FXI Studio a 5-second image-to-video clip costs from 10 credits, and free accounts can animate images with P-Video.
Start with 65 free credits. Creator plan $15/mo.
Try this prompt
“Bring this photo to life: a gentle breeze moves the hair and the grass, and the subject turns toward the light and smiles. Slow push-in, warm natural light, keep the face and setting exactly as in the image.”
Opens the studio with this prompt filled in · 65 free credits
Pick an image that can move
Image-to-video keeps everything the picture already shows: the framing, the light, the faces, the colors. So start with the frame you would want as the first frame of a shot. Leave room for motion: sky above a landscape, space in front of a person who will walk, a door that can open.
You can upload a photo you took or have permission to use, or draw the frame first on a still model in Create. Stills cost a few credits, and fixing the picture there is far cheaper than retaking video. Avoid images with small text or logos: the motion tends to smear them.
FXI Studio screens prompts and uploaded images before anything is generated. Use photos of people only with their permission.
Describe the motion, not the picture
The most common mistake is re-describing the image. The model can already see it, and new scene words fight the frame. Write three things instead: what moves in the scene, what the camera does, and how the shot ends.
"Fog drifts slowly between the pine trunks, a few birds lift off the branches, sunbeams shift as the clouds move. Slow push-in toward the man holding the photograph. Ends on a medium close-up, the photograph in focus."
Keep it to one camera move per clip. A push-in, a pan or an orbit reads as directed; three moves in five seconds reads as shaky. The prompt recipes page lists copy-ready wording for each camera move.
Choose a model for the shot
Every model in the table below takes your image as the first frame. P-Video is the fast default for camera moves and transitions, and it is the one free accounts can use. Seedance 2.0 Mini handles people and story shots, and Seedance 2.0 Fast gives sharper faces for the shots that matter most. Veo 3.1 Fast suits dialogue with native audio. Grok Imagine Video 1.5 costs 10 credits a clip but takes a start frame only.
On P-Video the attached image sets the aspect ratio, so crop the picture to the shape you want (16:9, 9:16 or 1:1) before you attach it.
Add a last frame to control where the shot lands
A first frame alone lets the model choose where the motion goes. Add a last frame and it builds the motion between two pictures you chose: a wide shot that pushes into a close-up, a room that turns into the next room, a day scene that ends at dusk.
This is also how you join clips without a cut. Pull the real last frame of a finished clip with the free last-frame extractor, use it as the first frame of the next clip, and the two play as one continuous shot.
Check the clip and fix it with motion words
Watch the clip once at full speed and once frame by frame. If the motion is too busy, cut adjectives and slow the camera. If faces drift, shorten the clip or move to a model built for people. If the shot ends somewhere you did not want, add a last frame or spell out the final framing in words.
Change one thing per retake so you can tell what fixed it. Most shots land in one or two tries once the prompt describes only the motion.
What it costs, and what is free
New accounts start with 65 free credits and no credit card, about 16 images or 3 free videos (2 today, 1 more next week), plus 22 more credits each week until the free balance is back at 65. Free clips run on P-Video at 5 seconds and 720p. Any credit pack or plan unlocks every model in the table, longer clips and 1080p.
On The Creator plan ($15/mo, 600 credits a month), a 5-second P-Video clip at 25 credits means about 24 clips a month.
| Model | Last frame too | Good for | Credits per 5 s clip | Free tier |
|---|---|---|---|---|
| P-Video | Yes | Camera moves and first-frame to last-frame transitions | 25 | Yes |
| Grok Imagine Video 1.5 | No | Cheap character motion from a start frame, with sound | 10 | Any pack or plan |
| LTX 2.5 Fast | Yes | Fast camera-controlled drafts | 20 | Any pack or plan |
| Seedance 2.0 Mini | Yes | Story shots with people, chained shot to shot | 25 | Any pack or plan |
| Seedance 2.0 Fast | Yes | Sharper faces and motion for hero shots | 75 | Any pack or plan |
| Veo 3.1 Fast | Yes | Dialogue shots with native audio | 90 | Any pack or plan |
Published per-unit prices for a 5-second clip at default settings; longer clips and 1080p cost more. Free accounts animate images with P-Video only: 3 free videos (2 today, 1 more next week), each 5 seconds at 720p. Any credit pack or plan unlocks every model in the table.
How we tested this
Model rows come from FXI Studio's Create picker: every model listed takes an image as its first frame, and the credit figures are the published per-model prices from the model benchmark page. Free-tier limits are the ones the server enforces. The example clip is a real P-Video shot from our kids counting-song tutorial.
Step by step
How to turn an image into a video with AI
- 1Pick or make the image
Upload a photo you have the rights to, or draw the frame on a still model in Create. Leave room in the frame for motion.
- 2Crop to the final shape
Crop to 16:9, 9:16 or 1:1 before attaching: on P-Video the image sets the aspect ratio.
- 3Attach it as the first frame
In Create, choose a video model from the table and drop the image in its first-frame slot.
Open Create - 4Write a motion-only prompt
Say what moves, what the camera does in one move, and how the shot ends. Do not re-describe the picture.
Camera move recipes - 5Add a last frame if needed
Attach a second image as the last frame when the shot has to land on a specific picture.
- 6Generate and check
Watch at full speed and frame by frame, then change one motion detail per retake.
- 7Chain the next shot
Extract the clip's real last frame and use it as the next clip's first frame for a continuous sequence.
Free last-frame extractor
Frequently asked questions
How do I turn a picture into a video with AI?
Attach the picture as the first frame of an image-to-video model and write a prompt that describes only the motion: what moves, the camera move and where the shot ends. In FXI Studio a 5-second clip costs from 10 credits, and free accounts can animate images with P-Video.
Is there a free AI image to video tool?
Yes, to start. New accounts start with 65 free credits and no credit card, about 16 images or 3 free videos (2 today, 1 more next week), plus 22 more credits each week until the free balance is back at 65. Free clips run on P-Video at 5 seconds and 720p.
What should I write in an image-to-video prompt?
Only the motion. Name what moves in the scene, give the camera one move such as a slow push-in or a pan, and say how the shot ends. The model already sees the image, so describing the scene again tends to fight the frame.
What is the difference between a first frame and a last frame?
The first frame is the picture the clip starts on. A last frame is a second picture the clip must end on, so the model builds the motion between the two. Use a last frame to control where a shot lands or to join two shots without a cut.
Can I animate an old family photo?
Yes. Upload it as the first frame and keep the motion gentle: a slow push-in, a small smile, hair moving in a breeze. Only use photos you took or have permission to use, of people who would be glad to see them move.
Why does my image-to-video clip look distorted?
Usually the prompt asks for too much motion or re-describes a different scene. Cut it to one camera move and one or two moving things, shorten the clip, and for close-ups of faces use a model built for people, such as Seedance 2.0 Mini or Fast.
How long can an image-to-video clip be?
It depends on the model: most clips are a few seconds to about 15 seconds, and P-Video goes up to 20 seconds on paid accounts. Free clips are 5 seconds. For longer scenes, chain clips by starting each one on the last frame of the one before.
Animate your image
Start with 65 free credits. Creator plan $15/mo.
Related guides
Hands-on tutorials
- TutorialHow to prompt P-Video for seamless transitions
- TutorialHow to turn a storyboard into one continuous shot
- TutorialHow to prompt Z-Image Turbo for film stills
Looking for a specific tool? Browse the documentation, the guides, or all tutorials.
